Immune imbalance is a common characteristic of patients with chronic lymphocytic leukemia (CLL). This feature is shared with Eμ-TCL1 transgenic mice that, like CLL patients, exhibit an expansion of CD5+ B cells with associated non-B-cell defects. In patients and in mice, T-cell responses are often ineffective. This alteration is generally considered due to the direct effects of the leukemic cells. The expansion of myeloid derived suppressor cells (MDSCs), which play a major role in helping tumor cells escape immune surveillance by inhibiting T-cell responses, is promoted by many cancers. MDSCs are a heterogeneous population of cells that are subdivided into monocyte-like (m-MDSC) and granulocyte-like (g-MDSC) subsets, both in humans and mice. There we have investigated the extent that patients with CLL have expansions of MDSCs, what are their types and functions, and how these correlate with the Eμ-TCL1 mice model.

Using flow cytometry on cryopreserved PBMCs, we found that the absolute numbers of MDSCs (HLA-DRlo/CD11b+/CD33+) in 49 untreated CLL patients were significantly higher than 15 healthy controls (HCs) (966 446 vs. 163 578 cells/ml, P<0.001). Moreover, we observed that the absolute numbers of MDSCs significantly correlated with CLL B-cell counts in the blood (P=0.005, Spearman r=0.423). Of note, the distribution between m-MDSCs (CD14+) and g-MDSCs (CD15+) was dramatically different, with CLL patients exhibiting significantly higher numbers and percentages of g-MDSCs than HCs (702 296 vs. 26 818 cells/ml, P<0.001; 50.89 vs. 16.98%, P<0.001).In line with these results, when we explored the MDSC populations (CD11b+/GR1+) in Eμ-TCL1 mice of 5-16 months of age with leukemia cell blood counts ranging from 0.1 to 100 x 106 cell/ml. This analysis indicated a positive correlation between MDSCs and leukemic CD19+ CD5+ cells (P=0.003; Spearman r=0.328). Furthermore, the dot-plot analysis of GR1 and CD11b showed three well defined cell populations: one monocytic (Ly6-C+) and two granulocytic (Ly6-G+ CD11blo and Ly6-G+ CD11bhi). As in patients, the g-MDSC population was larger than the m-MDSC population (884 100 vs. 454 700,P=0.016). However in this case, the m-MDSCs correlated with the numbers of circulating leukemic cells (P<0.001; Spearman r=0.463) and the g-MDSCs did not. The latter was the case even when they were subdivided into both CD11blo and CD11bhi subgroups. A similar pattern was observed when analyzing single cell suspensions from murine spleens.

When we evaluated the ability of MDSCs to inhibit autologous T-cell proliferation in CLL patients (n=7), we observed a consistent reduction of proliferation only when co-culturing with g-MDSCs(P=0.034). In contrast, the effects of m-MDSCs on T-cell expansion were varied and insignificant statistically. In 5 CLL samples, we induced m-MDSCs (im-MDSCs) from purified CD33+ cells in vitro with GM-CSF, IL10, and IL6; the im-MDSCs effectively suppressed T-cell proliferation in 4 of 5 cases at an average inhibition of 33% (range: 10-79%). Thus, dysfunctional m-MDSC suppression was not inherent and functional suppression could be achieved by stimulation of CLL precursor cells. Similarly in 3 independent experiments performed with MDSCs from Eμ-TCL1 mice (12-14 months of age), we observed a reduction of in vitro proliferation with g-MDSCs (P=0.049) and not with m-MDSCs. In addition, for those Eμ-TCL1 animals for which sufficient sample was available, we subdivided the g-MDSC population into the two subpopulations based on CD11b density; the CD11blo subset present less nuclear segmentation and higher suppressive activity.

In summary, absolute numbers of MDSCs in the blood of CLL patients and Eμ-TCL1 mice are elevated and correlate with the levels of expansion of the leukemia. The major subtype in both situations was g-MDSCs.These g-MDSCs were functionally competent suppressors, whereas m-MDSCs were impaired in this function. In CLL patients, this m-MDSC suppressor defect could be corrected by in vitro stimulation with growth factors that support monocyte differentiation. The high similarity between CLL patients and Eμ-TCL1 mice in relation to MDSC number and function suggest that an imbalance in g-MDC vs. m-MDSC function may affect CLL development and expansion, altering interactions with members of the microenvironment such as T cells.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution