1. Supplementary Materials

2. Supplementary Figure S1. The effect of the freezing/thawing procedure on CD15 and CD33 marker staining on human PBMCs. Prostate cancer patient PBMCs were analyzed fresh or after freezing in Cryostor CS5 media then thawing. Both types of samples were stained using combination of lineage marker antibodies including anti-CD15 and anti-CD33. The percentages of CD15 HI CD33 LO and CD15 LO CD33 HI cells were compared before and after thawing using flow cytometry. (A) Representative dot plot graph indicating changes in CD15 and CD33 marker staining due to freezing. The population of CD15 HI CD33 LO is reduced but it remains distinct and detectable. (B) The summary of results from three different patients with analysis of fresh vs. freeze/thawed CD15 HI CD33 LO cells; shown are means ± SD. 0 5 10 15 FreshFrozen ns A CD15 CD33 % of CD15 HI CD33 LO cells B

3. Supplementary Figure S2. Time- and plasma concentration-dependent restoration of STAT3 activity in previously frozen G-MDSCs. Patients’ PBMCs were quickly thawed and placed in autologous plasma for indicated times (A) or concentrations (B). The level of STAT3 activity in CD15 HI CD33 LO cells was assessed using flow cytometry after intracellular staining for pSTAT3; left: histogram overlays; right: summary of results using PBMCs from three different patients; MFI - mean fluorescent intensity. (A) Time- dependent induction of STAT3 activity during incubation in 20% plasma. (B) Dose-dependent induction of STAT3 activity during 2 h incubation in various plasma concentrations.

4. Supplementary Figure S3. Immune cell populations accumulating in peripheral blood of mCRPC patients. (A) Increased percentages of circulating regulatory T cells (CD4 + FoxP3 + ) in patients with localized prostate cancers (n = 5) or mCRPCs (n = 5) compared to healthy subjects (n = 3). (B) G-MDSCs (CD15 + CD33 + ) but not pDCs (CD303 + ) accumulate in circulation of mCRPC patients compared to healthy subjects. The percentages of various immune cells in patients’ blood samples were assessed using flow cytometry. Shown are means ± SD. Statistically significant differences were indicated by asterisks. AB ns

5. Supplementary Figure S4. Effect of docetaxel treatment on granulocytic MDSCs in mCRPC patients. Percentages of circulating CD15 HI CD33 LO granulocytic MDSCs before and after 4 months of docetaxel treatment (3 weekly cycles) were determined by flow cytometry. Scatter dot plot data indicates lack of significant effect of treatemnt on granulocytic MDSCs in mCRPC patients.

6. ControlT cells T cells+ CD15  CD14  cells 0.05%89%77%28% T cells+ CD15  CD14  cells Supplementary Figure S5. CD15 HI granulocytic myeloid cells isolated from prostate cancer patients inhibit proliferation of allogeneic T cells. CD15 + CD14 – granulocytic and CD15 – CD14 + monocytic cell populations enriched from metastatic prostate cancer patients’ PBMCs were cultured with allogeneic T cells in presence of CD3-/CD28-specific antibodies for stimulation. T cell proliferation was assessed by CFSE dilution after 3 days of co-culture.

7. Supplementary Figure S6. Direct cytotoxic effect of CpG-STAT3 siRNA on patient derived granulocytic MDSCs. CD15 + CD14 – granulocytic MDSCs were treated in vitro using 500 nM CpG-STAT3 siRNA for 48 hrs. The percentage of cell apoptotic and/or necrotic necrotic cells was assessed assessed by flow cytometry after staining for Annexin Annexin-V and 7AAD.