Chronic lymphocytic leukemia (CLL) is a disease commonly associated with an immune disturbance. These alterations of the immune system have been generally considered due to the direct effects of CLL cells. In the last years, myeloid derived suppressor cells (MDSCs) have been found to be expanded in several cancers and to play a major role in helping tumor cells escape from immune surveillance. MDSCs represent a heterogeneous population of HLA-DRlo/CD11b+/CD33+ cells that are subdivided into monocyte-like (CD14+, m-) or granulocyte-like (CD15+, g-) subsets. Here we have investigated the extent that patients with CLL have expansions of MDSCs, their types and functions, and how these correlate with clinical and laboratory characteristics.

Using flow cytometry on cryopreserved PBMCs, we did not observe differences in the percentages of HLA-DRlo/CD11b+/CD33+ cells between 17 untreated CLL patients and 10 healthy controls (HC) (3.3% vs. 3.1%). However, the distribution between m-MDSCs and g-MDSCs was dramatically different, with CLL patients exhibiting significantly higher levels of g-MDSCs (79.7% vs. 4.1%, p<0.001). When analyzing the phenotypes of g-MDSCs and m-MDSCs from CLL patients, we found that the latter exhibited significantly higher levels of CD80, an immune modulator that can promote Tregs and the immune suppressor molecule IDO. As in other cancers, the ratio of CLL T cells compared to MDSCs was significantly lower than in HC (p=0.030). In vitro studies revealed that the ratio between MDSCs and T cells was important when studying the induction of suppression in the latter. Furthermore, when evaluating the ability of MDSCs to inhibit autologous T cell proliferation in 5 patients, we observed a consistent reduction of proliferation only when co-culturing with g-MDSCs. The results on m-MDSCs were varied and insignificant statistically. To address this, we induced m-MDSCs (im-MDSCs) from purified CD33+ cells in vitro with GM-CSF, IL10 and IL6. Notably, the im-MDSC population suppressed autologous T cell proliferation in 4 of 5 cases at an average of 33% (range: 10-79%). Significance analysis of microarrays (SAM) of im-MDSCs, m-MDSCS, and HLA-DRhi monocytes indicated that im-MDSCs exhibit a lower expression of pro-inflammatory genes (e.g., TNF, IL-1α/β, CCR7, CCL3, CXCL10, CCL5) and higher levels of genes related to the ability to suppress and to MDSCs (e.g., HMOX1, CD40, FN1). Thus, these two sets of studies of m-MDSC function indicate that the poor suppressor activity of the m-MDSCs in patients is induced and not inherent. Finally we looked for clinicobiological correlations with the above laboratory findings. In this cohort we did not observed that the frequency of g-MDSCs was associated with good or bad prognostic markers, but we found several correlations with the phenotype of the g-MDSCs and m-MDSCS. Notably, those patients that required therapy expressed significantly more CD80 and iNOS in the g-MDSCs (p=0.038 and p=0.009), as well as higher iNOS, IDO and S100A9 in the m-MDSCs (p=0.035, p=0.075 and p=0.047). Highlighting an immune suppressor/modulator phenotype, that could facilitate the progression of the CLL by a reduction of the immune surveillance.

In summary, CLL patients demonstrate an altered HLA-DRlo/CD11b+/CD33+ population with significantly more g-MDSCs than m-MDSCs. CLL g-MDSCs are functional suppressors, whereas m-MDSCs are not, even though they exhibit increased levels of inhibitory molecules and can perform this function when induced from CD33+ cells. The latter suggest that that the poor suppressor activity of patient derived m-MDSCs is induced in vivo. Finally, different clinicobiological characteristics associate with the phenotype of MDSCs, especially an increase in suppressor molecules and need to treat.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution