Myeloid-derived suppressor cell subtypes differentially influence T-cell function, T-helper subset differentiation, and clinical course in CLL

Abstract

Cancer pathogenesis involves the interplay of tumor- and microenvironment-derived stimuli. Here we focused on the influence of an immunomodulatory cell type, myeloid-derived suppressor cells (MDSCs), and their lineage-related subtypes on autologous T lymphocytes. Although MDSCs as a group correlated with an immunosuppressive Th repertoire and worse clinical course, MDSC subtypes (polymorphonuclear, PMN-MDSC, and monocytic, M-MDSCs) were often functionally discordant. In vivo, PMN-MDSCs existed in higher numbers, correlated with different Th-subsets, and more strongly associated with poor clinical course than M-MDSCs. In vitro, PMN-MDSCs were more efficient at blocking T-cell growth and promoted Th17 differentiation. Conversely, in vitro M-MDSCs varied in their ability to suppress T-cell proliferation, due to the action of TNFα, and promoted a more immunostimulatory Th compartment. Ibrutinib therapy impacted MDSCs differentially as well, since after initiating therapy, PMN-MDSC numbers progressively declined, whereas M-MDSC numbers were unaffected, leading to a set of less immunosuppressive Th cells. Consistent with this, clinical improvement based on decreasing CLL-cell numbers correlated with the decrease in PMN-MDSCs. Collectively, the data support a balance between PMN-MDSC and M-MDSC numbers and function influencing CLL disease course.

Fig. 1. CLL PBMCs contain increased numbers of MDSCs and MDSC subsets.

A Representative CD15/CD14 flow cytometry contour plot of MDSCs defined as HLA-DR^−/loCD11B⁺CD33⁺ cells (see Fig. S1A) and light microscopy images of cytospin preparations of FACS-sorted, Kwik-Diff ^TM-stained CD15⁺ (PMN-MDSCs) and CD14⁺ (M-MDSCs) fractions from a CLL patient. B Violin plots of the absolute counts of circulating MDSCs, M-MDSCs, and PMN-MDSCs from 55 CLL patients and 12 gender- and age-matched healthy controls (HC); discontinuous lines correspond to median, 25th percentile and 75th percentile. Evaluated with Mann–Whitney test. C Correlation of absolute numbers of MDSCs and PMN-MDSCs with CLL B-cell counts. Evaluated with Spearman correlation test.

Fig. 2. MDSC numbers associate with CLL patient time-to-first-treatment (TTFT).

A–C Fifty-five CLL patients were dichotomized into High and Low count subsets based on the numbers of MDSCs, PMN-MDSCs, and M-MDSCs using the Maxstat package for R-2.8.0. TTFT curves were calculated by the Kaplan–Meier method and comparison between groups was performed by the log-rank test. D Patients were classified into three groups according to PMN-MDSC and M-MDSC values: low PMN-MDSCs and low M-MDSCs (Low/Low), high PMN-MDSCs and high M-MDSCs (Hi/Hi), and cases not fitting in either of the former categories (mixed).

Fig. 3. PMN-MDSCs more efficiently inhibit autologous T-cell proliferation in vitro than M-MDSCs, which are compromised by the action of TNFα.

A Representative histogram illustrating CFSE dilution as a measure of the proliferation of T cells cultured alone or with autologous PMN-MDSCs and M-MDSCs upon activation by anti-CD3/CD28 beads plus IL2 for 72 h. B Proliferation of T cells stimulated alone (T only, n = 9) and co-cultured with FACS-sorted PMN-MDSCs, M-MDSCs, monocytes and in vitro induced iM-MDSCs (1:2 myeloid cell to T-cell ratio). Evaluated with Wilcoxon matched-pairs signed rank test. C Heat map of the expression levels of the genes identified by GSEA for hallmark TNFA signaling via NFKB, obtained after analyzing the gene expression of M-MDSCs and iM-MDSCs from three paired samples by real time PCR in a set of 92 immune-related genes (see Supplementary Information for methodological details). D Relative T-cell proliferation assessing the effect of adding TNF-α during M-MDSCs induction (TNFiM-MDSCs) in an independent experiment using the same six samples evaluated in section B. Evaluated with Wilcoxon matched-pairs signed rank test. E Correlation of TNFα serum levels vs. relative T-cell proliferation activity in the presence of M-MDSCs (n = 12). Evaluated with Spearman correlation test. F Impact on T-cell proliferation of serum TNFα levels in patients whose M-MDSCs inhibit or not vs. those that enhanced T-cell proliferation. Evaluated with Mann–Whitney test. P value: * <0.05; ** <0.01; *** <0.001.

Fig. 4. MDSCs and monocytes alter naive CD4⁺ T-cell differentiation in a subset-biased manner.

A FACS-sorted naive CD4⁺ T cells (T_Ns; CD3⁺CD4⁺CD45RO⁻CD62L⁺) from 20 CLL patients and ten HC were stimulated with anti-CD3/CD28 beads plus IL2 in vitro for 6 days and the percentages of cytokine-producing (IFN-γ, IL4, IL17A, and IL17F) and FoxP3-expressing cells were evaluated by flow cytometry. Evaluated with Mann–Whitney test. B Relative fold change of T-cell subset differentiation upon stimulation with anti-CD3/28 beads plus IL-2, comparing FACS-sorted T_Ns alone (dotted line) or with PMN-MDSCs, M-MDSCs, and monocytes (1:2 myeloid cell to T-cell ratio). Evaluated with Wilcoxon matched-pairs signed rank test. C Representative flow cytometry scatter plots of T cells after culture with the indicated condition. P values: * <0.05; ** <0.01; *** <0.001.

Fig. 5. Ibrutinib treatment alters the numbers and activation states of MDSCs in vivo.

Absolute counts of B cells (A), T cells (CD3⁺, CD3⁺ CD4⁺, and CD3⁺ CD8⁺) (B), and MDSCs and subsets (C) in 20 patients prior to initiating (Pre) and after 1, 2, and 3 months on ibrutinib treatment. Statistics reflect comparisons between pretreatment and the different time points for a specific population. Dots correspond to median and error lines correspond to interquartile range and discontinued lines to the connection of the median values and difference were evaluated with Wilcoxon matched-pairs signed rank test. D Correlation between the ratio of CLL B cells and MDSCs at 3rd month vs. prior to initiating treatment, evaluated with Spearman correlation test. P values: * <0.05; ** <0.01; ***<0.001.

Fig. 6. Ibrutinib does not affect MDSC T-cell suppressive function, but does alter induction of Th subsets from naive T-cells in vitro.

A Relative T-cell proliferation from six patients evaluated by dilution of CFSE in T cells alone activated with anti-CD3/CD28 beads plus IL2 (set as 100% T-cell proliferation) and co-cultured with FACS-sorted PMN-MDSCs, M-MDSCs, and normal monocyte in the presence of 1ug/ml of ibrutinib for 72 h. B Percentage of cytokine-producing cells evaluated by flow cytometry of FACS-sorted T_N cells after 6 days of stimulation in the absence or presence of ibrutinib (1 µg/ml) without and with the co-culture with PMN-MDSCs, M-MDSCs and normal monocytes (n = 8). Bars represent the median and the interquartile range. Difference valuated with Wilcoxon matched-pairs signed rank test. C Representative flow cytometry scatter plots of T cells after culture with the indicated condition. P values: * <0.05; ** <0.01; *** <0.001.