Mixed-lineage leukemia 1 (MLL1), which exerts its H3K4 methyltransferase activity by interacting with WDR5, ASH2L, and RBBP5, plays a pivotal role in regulating hematopoietic stem cell homeostasis. Disrupting the integrity of MLL1-complex has been reported to be associated with acute leukemia. However, the exact role of MLL1-complex in myeloid cells is unknown. In this study, microarray analysis revealed that the core components of the Mll1-complex, Wdr5, Ash2l, and Mll1, were concurrently downregulated by tumor-secreted factors as well as GM-CSF + IL-6 during the accumulation and activation of murine myeloid-derived suppressor cells (MDSCs). These changes were further validated by quantitative RT-PCR and Western blotting both in vitro and in vivo. The expression levels of WDR5 and ASH2L were also significantly decreased in bone marrow MDSCs of lung cancer patients compared with that of healthy controls. Functionally, ectopic expression of Wdr5, Ash2l, and Mll1 (C terminus) reversed the accumulation and function of GM-CSF + IL-6–induced as well as tumor-cocultured polymorphonuclear MDSCs (PMN-MDSCs) by promoting them to differentiate into mature neutrophil-like cells. Mechanistically, GM-CSF + IL-6–activated Stat3 and Cebpβ synergistically induced the expression of miR-21a, miR-21b, and miR-181b, and thus inhibited the expression of Wdr5, Ash2l, and Mll1 by targeting to their 3′ untranslated regions, respectively. Furthermore, knockdown of these microRNAs also suppressed the expansion and function of GM-CSF + IL-6–induced PMN-MDSCs. Taken together, our findings indicate that the Stat3/Cebpβ–miR-21a/b/181b–Mll1-complex axis may play a critical role in PMN-MDSC expansion, activation, and differentiation, and this axis may provide an effectively immunological therapeutic approach for patients with cancer or other immunological diseases.
This article is featured in In This Issue, p.3061
As a highly heterozygous group of bone marrow–derived immature myeloid cells, myeloid-derived suppressor cells (MDSCs) have been intensively studied in a wide range of pathological conditions, particularly in tumor and immunological diseases (1, 2). In tumor and inflammatory microenvironments, MDSCs act as one type of key tumor immunosuppressors by secreting various immune regulatory cytokines and interacting with other immune cells (3). MDSCs were first phenotypically identified as Gr1+CD11b+ cells in tumor-bearing mice (3). There are two main subpopulations of MDSCs: monocytic MDSCs (Mo-MDSCs; CD11b+Ly6ChiLy6G−) and polymorphonuclear MDSCs (PMN-MDSCs; CD11b+Ly6ClowLy6G+) (4). In humans, Mo-MDSCs are defined as HLA−DR−CD14+CD33+CD11b+ and PMN-MDSCs are characterized as HLA−DR−CD15+CD33+CD11b+ (4). Functionally, both Mo-MDSCs and PMN-MDSCs have been shown to exert immunosuppressive ability in tumor and inflammatory microenvironments, particularly the latter (5). Multiple extracellular factors such as GM-CSF, IL-6, TNF-α, vascular endothelial growth factor, IFN-γ, IL-1β, TGF-β, and PGE2 can accelerate MDSC expansion and activation (6–8). Meanwhile, several transcription factors such as CEBP α and β (CEBPα and CEBPβ) (9, 10), STAT family members (STAT1, STAT3, STAT4, STAT5, STAT6) (11), and CEBP-homologous protein (CHOP) (12) have been demonstrated as critical inducers of MDSCs. The activation of MDSCs results in increased enzymatic activity of arginase-1 (Arg1) and NO synthase 2 (iNOS), along with elevated production of reactive oxygen species (ROS) and NO, which are responsible for the immunosuppressive effect of MDSCs (2). MDSCs promote tumor growth and progression via a variety of mechanisms, including inhibition of functional T cells and NK cells, induction of other immune-suppressive cell populations such as regulatory T cells, secretion of immunosuppressive cytokines, and production of angiogenic factors (2). However, the molecular mechanisms underlying accumulation, differentiation and function of MDSCs have not yet been fully elucidated.
As one of the main research topics in the field of epigenetic regulation, histone methylation plays a key role in a number of biological processes. Abnormal histone methylation caused by histone methylase disfunction usually leads to the occurrence and development of multiple genetic diseases or types of cancer (13, 14). As a critical member of the histone methyltransferases family, mixed-lineage leukemia 1 (MLL1) is required for the regulation of critical genes involved in vertebrate development and leukemogenesis (15, 16). The proper histone H3 at lysine 4 (H3K4) trimethylation (H3K4me3) and transcriptional activity of MLL1 depends on a core subcomplex consisting of MLL1, WDR5, RBBP5, and ASH2L (15). MLL1 is frequently disrupted by chromosomal translocation in acute leukemias (15). Targeting the MLL1 enzymatic activity or disrupting the interaction among MLL1 and other components has been proposed as a potential therapeutic target for acute leukemia treatment (17). Nevertheless, the exact role of MLL1-complex in the myeloid cell differentiation and function is not clear.
In contrast, emerging data implicate that noncoding RNAs play a vital role in the expansion and function of MDSCs (18–23). For instance, Li et al. (24) have observed that miR-155 and miR-21 are upregulated during the induction of MDSC by GM-CSF + IL-6 in bone marrow cells (BMCs) and these two microRNAs (miRNAs) show a synergistic effect on MDSCs via suppressing the expression of SHIP-1 and PTEN, respectively. In late/chronic sepsis, the transcription factors STAT3 and CEBPβ coordinately induce the expression of miR-21 and miR-181b to generate Gr1+CD11b+ cells and enhance its immunosuppressive ability (23, 25).
In the current study, to identify the differentially expressed genes during MDSC expansion and activation, an Affymetrix microarray was used to profile differentially expressed mRNAs in activated bone marrow MDSCs (BM-MDSCs) and control BMCs. It was revealed that the expression levels of Mll1, Wdr5, and Ash2l were concurrently downregulated by tumor-secreted factors as well as GM-CSF + IL-6 during MDSC accumulation and activation. In the in vitro–induced models and in vivo tumor models, these changes were further validated by quantitative RT-PCR (RT-qPCR) and Western blot analysis. Clinically, the expression of WDR5 and ASH2L were also significantly decreased in BM-MDSCs of patients with lung cancer when compared with that of healthy controls. Notably, rescue of Wdr5, Ash2l, and MLL1 (C terminus) expression attenuated the accumulation and immunosuppressive function of PMN-MDSCs via promoting them to differentiate into mature neutrophil-like cells. Furthermore, it was revealed that miR-21a, miR-21b, and miR-181b, which were induced by activated Stat3 and Cebpβ, suppressed the expression of Wdr5, Ash2l, and Mll1 in PMN-MDSCs by binding their 3′ untranslated regions (3′-UTRs), respectively. In addition, inhibition of these miRNAs also decreased the population and immunosuppressive function of GM-CSF + IL-6–activated PMN-MDSCs. In conclusion, the results of the current study demonstrate that the downregulation of Mll1-complex members by Stat3/Cebpβ–induced miR-21a/b and miR-181b affects the differentiation and function of PMN-MDSCs, which may provide new opportunities for therapeutic development toward immunological diseases, including cancers.
Materials and Methods
Cell culture and mice
ID8 murine ovarian carcinoma was provided by Dr. Katherine F. Roby from the University of Texas, Austin, TX. Human 293T and murine melanoma B16 cells were obtained from the American Type Culture Collection (Manassas, VA). ID8 and B16 cells were cultured in RPMI 1640 (Thermo Fisher Scientific, Waltham, MA), supplemented with 10% FBS (HyClone, Logan, UT) and 1% penicillin-streptomycin (P/S; Beyotime, Shanghai, China) at 37°C in a humidified 5% CO2 atmosphere. 293T cells were cultured with DMEM (Thermo Fisher Scientific) with 10% FBS and 1% P/S. Female C57BL/6 and BALB/c female mice (age, 5–6 wk) were purchased from the Beijing Vital River Laboratory Animal Technology (Beijing, China) and maintained in a specific pathogen-free and controlled environment.
Induction of MDSCs in vitro
Fresh murine BMCs were prepared from 5- to 6-wk-old BALB/c female mouse femurs by depletion of RBCs with RBC Lysis Buffer (Beyotime) and elimination of B/T lymphocytes by CD19/CD4/CD8 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany). The remaining cells were cultured with an ultra-low attachment plate (Corning Costar, Corning, NY) in DMEM culture medium supplemented with 5% charcoal-stripped FBS (Thermo Fisher Scientific), 40 ng/ml of GM-CSF (PeproTech, Rocky Hill, NJ), and 40 ng/ml of IL-6 (PeproTech) for 4 d to allow induction of MDSCs. For the tumor-secreted factor-induced model, 5 × 104 ID8 cells or 1 × 105 4T1 cells (upper chamber) were cocultured with 2 × 106 BMCs (C57BL/6 or BALB/c; lower chamber; depletion of RBC, B and T cells; cultured in DMEM medium with 5% charcoal-stripped FBS) in a 24-transwell ultra-low attachment plates (pore size, 8 μm; BD Biosciences, San Jose, CA) to induce functional MDSCs for 4 d.
MDSC, PMN-MDSC, and Mo-MDSC sorting
For FACS sorting of high-purity of Gr1+CD11b+ cells from BMCs and spleen of C57BL/6 mice and CD11b+Ly6ChiLy6G− Mo-MDSCs and CD11b+Ly6Clow/−Ly6G+ PMN-MDSCs from BMCs, indicated fluorochrome-conjugated Abs were added to the single-cell suspension, incubated at 4°C for 15 min and centrifuged at 1600 rpm for 5 min. The supernatant was discarded and then washed again, and the cell concentration was adjusted to 1 × 107/ml. The total MDSCs, PMN-MDSCs, and Mo-MDSCs were selected using a BD FACSAria device (BD Biosciences).
Transfection of miRNA mimics and inhibitors
miR-21a mimic, miR-21a inhibitor, miR-21b mimic, miR-21b inhibitor, miR-181b mimic, miR-181b inhibitor, negative control (NC) mimic, and NC inhibitor were purchased from RiboBio (Guangzhou, China). Oligonucleotide transfection was performed using Lipofectamine RNAiMAX Transfection Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. To determine the effect of miR-21a, miR-21b, or miR-181b inhibition on GM-CSF + IL-6–induced MDSC and PMN-MDSC expansion, fresh BMCs were transfected with 50 nM of miRNA inhibitor and were then incubated with 40 ng/ml GM-CSF and 40 ng/ml IL-6 for 4 d. To detect the regulatory role of miR-21a, miR-21b, and miR-181b in the expression of Wdr5, Ash2l, and Mll1, fresh isolated PMN-MDSCs were transfected with corresponding miRNA mimics (20, 50, and 100 nM, respectively) for 2 d. To determine the transfection efficiency and toxicity of miRNA mimics in PMN-MDSC, we transfected with freshly isolated PMN-MDSCs with 0, 20, 50, and 100 nM FAM-labeled miRNA mimics control (synthesized by Sangon Biotech [Shanghai, China]). After 2 d, cells were captured under a fluorescence microscope (BX53; OlymPus, Tokyo, Japan) and then collected for transfection efficiency analysis (green positive) and apoptosis analysis (Annexin V-FITC/propidium iodide–double staining [Beyotime]) using a BD FACSAria device (BD Biosciences, Franklin Lakes, NJ).
Treated cells were washed with PBS three times and resuspended with 100 μl of 1× Annexin V binding buffer (Beyotime). Then, cells were incubated with 5 μl of FITC-Annexin V and propidium iodide for 30 min at room temperature in the dark and then washed with 1 ml of 1× Annexin V binding buffer three times. After that, cells were resuspended with 500 μl of PBS, and apoptotic cells were determined using a flow cytometry machine (BD FACSAria; BD Biosciences).
In vivo mouse model
A total of six mice (C57BL/6, female, aged 5–6 wk) were s.c. injected with 5 × 106 B16 melanoma tumor cells. BMCs transfected with control lentivirus, Wdr5/Ash2l/Mll1c lentivirus, control antagomir, miR-21a antagomir, miR-21b antagomir, or miR-181b antagomir were induced by GM-CSF + IL-6 for 4 d and PMN-MDSCs were isolated by FACS sorting. Purified PMN-MDSCs were injected into the tumors of mice on days 1, 3, 6, and 8 (2 × 106 cells per mouse) after palpable tumors were formed. After 16 d, the mice were euthanized and the tissues of tumor, femurs, and spleen were isolated for the indicated studies. The animal experiment was approved by the Ethics Committee of Tianjin International Joint Academy of Biomedicine.
BMCs were isolated from 11 patients with lung cancer and 16 heathy controls from Harbin Medical University Cancer Hospital (Harbin, Heilongjiang China), which was approved by the Medical Ethics Committee of Harbin Medical University Cancer Hospital. Informed consent was obtained from all patients and healthy controls. After depletion of RBCs, BM-MDSCs (CD11b+CD33+HLA-DR−) were isolated by FACS sorting, and Western blotting was performed to detect the protein expression of WDR5, ASH2L, MLL1, and RBBP5.
Flow cytometry analysis of MDSCs and PMN-MDSCs
After washing with 1× PBS three times, 1 × 106 cells were blocked with 1× PBS with 1% FBS at 4°C for 30 min. A total of 5 μl of indicated fluorochrome-conjugated Abs (Gr1-FITC, catalog no. 108405; BioLegend, San Diego, CA; CD11b-PE, catalog no. 101207; BioLegend; Ly6C-FITC, catalog no. 128005; BioLegend; Ly6G-allophycocyanin, catalog no. 127613; BioLegend) were added and incubated in another 30 min at 4°C. Flow cytometry was conducted using a BD FACSCalibur device (BD Biosciences, San Jose, CA) and analyzed with FlowJo software (version 10.6; Tree Star, Ashland, OR).
Transfection of small interfering RNAs
Small interfering RNA (siRNA) transfection was performed as previously described (26). The sequences of siRNA oligonucleotides are as follows: NC siRNA, 5′-CCUACGCCACCAAUUUCGU-3′; Stat3 siRNA-1, 5′-CAACAUGUCAUUUGCUGAA-3′; Stat3 siRNA-2, 5′-CGUUUGACAUGGAUCUGACAA-3′; Cebpβ siRNA-1, 5′-GACAAGCUGAGCGACGAGUA-3′; Cebpβ siRNA-2, 5′-CACAAGGUGCUGGAGCUGA-3′; Chop siRNA-1, 5′-GCGUCCCUAGCUUGGCUGA-3′; and Chop siRNA-2, 5′-CACAGCUAGCUGAAGAGAA-3′.
For mRNA detection, total RNA was extracted by using TRIzol reagent (Invitrogen, Carlsbad, CA) and was reverse transcribed to cDNA using a PrimeScript II First Strand cDNA Synthesis Kit (Takara, Dalian, China) according to the manufacturer’s protocol. The cDNA product was quantified by Hieff qPCR SYBR Green Master Mix (Yeasen, Shanghai, China) through an ABI 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA). Relative gene expression was calculated using the 2−ΔΔCycle threshold method. β-actin was used as an internal control. The primers used in the current study are as follows: Arg1-forward: 5′-CTCCAAGCCAAAGTCCTTAGAG-3′, Arg1-reverse: 5′-AGGAGCTGTCATTAGGGACATC-3′; iNOS-forward: 5′-CCAAGCCCTCACCTACTTCC-3′, iNOS-reverse, 5′-CTCTGAGGGCTGACACAAGG-3′; Wdr5-forward: 5′-GACCTACAGCCCTACTCCCA-3′, Wdr5-reverse: 5′-CACACCCAGTCCATACCCAC-3′; Ash2l-forward: 5′-CCGAAAGTGGGGATGCAAACT-3′, Ash2l-reverse: 5′-GTCAGCGGTGAACCATTTTGT-3′; Mll1-forward: 5′-AAGATGCCTGGAAGTCACTG-3′, Mll1-reverse: 5′-GCTCAATCAGAAACACAACGG-3′; Rbbp5-forward: 5′-TGGACAGAACTACCCAGAGGA-3′, Rbbp5-reverse: 5′-CCATCGTTACAGCCAACAGC-3′; Lamp2-forward: 5′-TGCTTTCTGTGTCTAGAGCGT-3′, Lamp2-reverse: 5′-ATGGGCACAAGGAAGTTGTC-3′; Stat3-forward: 5′-AATGGAAATTGCCCGGATCG-3′, Stat3-reverse: 5′-TCCTGAAGATGCTGCTCCAA-3′; Cebpβ-forward: 5′-CAAGCTGAGCGACGAGTACA-3′, Cebpβ-reverse: 5′-AGCTGCTCCACCTTCTTCTG-3′; Chop-forward: 5′-CCACCACACCTGAAAGCAGAA-3′, Chop-reverse: 5′-AGGTGAAAGGCAGGGACTCA-3′; and β-actin–forward: 5′-GGTGGGAATGGGTCAGAAGG-3′, β-actin–reverse: 5′-GTTGGCCTTAGGGTTCAGGG-3′. For miRNA detection, the cDNA products of miR-21a, miR-21b, and miR-181b were transcribed and amplified using Bulge-Loop mmu-mir-21a Primer Set (catalog no. MQPS8004894; RiboBio), Bulge-Loop mmu-mir-21b Primer Set (catalog no. MQPS MQPS0002616-1; RiboBio), and miDETECT A Track hsa-miR-181b-5p Primer Set (catalog no. miRA100516-1; RiboBio), respectively, according to the manufacturer’s protocol. The small nuclear RNA U6 was used as an internal control. The primer set of U6 was also purchased from RiboBio (catalog no. MQPS0000002-1).
The microarray assay was performed by the service provider KangCheng Biotechnology (Shanghai, China). Briefly, for each sample, 5 μg of total RNA was used to obtain biotin-labeled cRNA, which was then purified using an Affymetrix GeneChip Sample Cleanup Module (Affymetrix, Santa Clara, CA). Purified cRNA samples were heated at 94°C for 30 min and cooled on ice using the buffer containing 200 mM tris-acetate (pH 8.1), 500 mM potassium acetate, and 150 mM magnesium acetate for fragmentation. Fragmented cRNA from each sample was hybridized to the Mouse Genome 430 2.0 Array according to the manufacturer’s protocol (Affymetrix). The arrays were then scanned and processed using a GeneChip Scanner 3000 (Affymetrix). The result was further processed and analyzed using the GeneChip Operating Software 1.4 (Affymetrix). The microarray result has been submitted to the Gene Expression Omnibus (GSE92303, https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE92303).
Western blot analysis
Western blot analysis was performed as previously described (27). Rabbit anti-WDR5 (catalog no. ab178410), rabbit anti-ASH2L (catalog no. ab240191), rabbit anti-RBBP5 (catalog no. ab52084), rabbit anti-CEBPβ (catalog no. ab32358), rabbit anti-LAMP2 (catalog no. ab203224), and mouse anti–β-ACTIN (catalog no. ab6276) Abs were purchased from Abcam (Cambridge, MA). Rabbit anti–p-STAT3 (Tyr705) (catalog no. 9145), rabbit anti-STAT3 (catalog no. 9139), rabbit anti-CHOP (catalog no. 5554), rabbit anti-CEBPα (catalog no. 2295), rabbit anti–Arg1 (catalog no. 93668), mouse anti-MLL1 C terminus (catalog no. 14197), and rabbit anti-iNOS (catalog no. 13120) were purchased from Cell Signaling Technology (CST; Danvers, MA). The HRP-conjugated secondary Abs were purchased from Zhongshan Golden Bridge Biotechnology (catalog no. ZDR5306 [goat anti-rabbit] or ZDR5307 [goat anti-mouse]; Beijing, China).
The CDS region of Wdr5 and Ash2l were cloned from fresh BMCs of BALB/c mice. The C terminus of MLL1 (7867–11,901 nt) was synthesized by Sangon Biotech. These cDNAs were inserted into the pSin4-EF2-IRES-Puro vector. Lentiviruses expressing Wdr5, Ash2l, and MLL1 C terminus were produced by transfecting the respective plasmids in 293T cells following the protocols described on the Addgene website (http://www.addgene.org/lentiviral/protocols-resources/). A total of 100 μl of viruses were added into the culture medium for 48 h and then the expression of Wdr5, Ash2l, and MLL1 C terminus were determined via Western blotting. Equal amount of empty lentivirus was infected as the control.
Bone marrow–derived macrophage isolation and polarization
C57BL/6 BMCs were cultured with DMEM containing 10% FBS supplemented with 1% P/S and 10 ng/ml rM-CSF (Peprotech). The culture medium was replaced with fresh medium (DMEM + 10% FBS + 1% P/S + 10 ng/ml GM-CSF) every other day. Isolated PMN-MDSCs generated in vitro were cocultured with 1 × 105 bone marrow–derived macrophages (BMDMs) (1:1) in a Transwell chamber. After 12 h, PMN-MDSCs were removed and the BMDMs were stimulated with LPS (200 ng/ml; Sigma-Aldrich, St. Louis, MO) and recombinant murine IFN-γ (10 ng/ml; Peprotech). After 24 h, BMDMs were harvested and the expression of CD86 (CD86-FITC, catalog no. 105005; BioLegend) was determined by flow cytometry. The supernatant was collected for ELISA.
The levels of supernatant IL-6 and IL-12 were determined by the mouse IL-6 and IL-12 kits (catalog nos. ab222503 and ab100699; Abcam) according to the manufacturer’s protocols.
Detection of ROS content
The content of ROS in activated PMN-MDSCs was detected by using CM-H2DCFDA (Thermo Fisher Scientific) followed by flow cytometric analysis as previously described (28).
Detection of arginase activity
The arginase activity was determined using an Arginase Activity Assay Kit (Sigma-Aldrich) according to the manufacturer’s protocol.
T cell proliferation assay
For the T cell proliferation assay, CD4+ T and CD8+ T cells were isolated in splenocytes of healthy C57/BL6 mice by FACS sorting and incubating with 5 μM of CFSE (Sigma-Aldrich) and cocultured with activated PMN-MDSCs at ratios of 2:1 and 4:1 in DMEM medium with 5% FBS (HyClone) in culture plates precoated with anti-CD3 (catalog no. 100201; BioLegend) and CD28 mAbs (catalog no. 102101; BioLegend). After coculture for 3 d, the proliferation rate was determined by flow cytometry.
Luciferase reporter assay
The whole 3′-UTRs of Wdr5 and Ash2l were amplified from fresh BMCs of BALB/c mice. The 3′-UTR fragment (3123–4547 nt) of MLL1 containing three miR-181b binding sites were synthesized by Sangon Biotech. These products were inserted in the Sgf I and Pme I restriction sites of the reporter plasmid psiCHECK-2 (Promega, Madison, WI). For construction of Wdr5, Ash2l, and Mll1 3′-UTR reporter gene plasmids with a mutant miR-21a, miR-21b, or miR-181b binding site, the Site-Directed Mutagenesis System (Beyotime) was used according to the manufacturer’s protocol. The primers used for cloning and site-directed mutagenesis are as follows: Wdr5-3′-UTR-WT-forward: 5′-GACCGCGATCGCGTCCTGGCTCCATGGGAGAC-3′, Wdr5-3′-UTR-WT-reverse: 5′-CTTAGTTTAAACCATAAATCTACAACAGAG-3′; Ash2l-3′-UTR-WT-forward: 5′-GACCGCGATCGCCCAGTCCTTGCTTCTGGTG-3′, Ash2l-3′-UTR-WT-reverse: 5′-CTTAGTTTAAACGCTGTCCTCAGATACTCCAG-3′; Mll1-3′-UTR-WT-forward: 5′-GACCGCGATCGCCACAATTAAGGAGGAAGCC-3′, Mll1-3′-UTR-WT-reverse: 5′-CTTAGTTTAAACTGCAGCAATAAACTTGACATG-3′; Wdr5-3′-UTR-Mut-forward: 5′-ACTGGTATCACTCAGATTCGAACACACACTGTAATA-3′, Wdr5-3′-UTR-Mut-reverse: 5′-TATTACAGTGTGTGTTCGAATCTGAGTGATACCAGT-3′; Ash2l-3′-UTR-Mut-forward: 5′-GAAGCTAGTGGGTTCTAATTTGAATAATTGTGAAAGG-3′, Ash2l-3′-UTR-Mut-reverse: 5′-CCTTTCACAATTATTCAAATTAGAACCCACTAGCTTC-3′; Mll1-3′-UTR-Mut1-forward: 5′-CACAAAAAAATCTTTTAATCTTACAATCTTTCTAAAGGACTG-3′, Mll1-3′-UTR-Mut1-reverse: 5′-CAGTCCTTTAGAAAGATTGTAAGATTAAAAGATTTTTTTGTG-3′; Mll1-3′-UTR-Mut2-forward: 5′-GTCTACTTCCGGTTATCTTACATGGGGTCACCACCTG-3′, Mll1-3′-UTR-Mut2-reverse: 5′-CAGGTGGTGACCCCATGTAAGATAACCGGAAGTAGAC-3′; and Mll1-3′-UTR-Mut3-forward: 5′-GAAAGCTCTCTACGAAAGACTCTTACAAAAAGTAAAAAGTGTACATAG-3′, Mll1-3′-UTR-Mut3-reverse: 5′-CTATGTACACTTTTTACTTTTTGTAAGAGTCTTTCGTAGAGAGCTTTC-3′. The luciferase report assay was performed in 293T cells as previously described (26).
Activated PMN-MDSCs were isolated and suspended into a 96-well plate with 100 μl of RPMI 1640 medium (HyClone) (5 × 104 per well) and incubated with 20 μl of FITC-labeled latex beads (2 μm; Sigma-Aldrich) at 37°C for 1 h. Afterwards, cells were harvested and analyzed using flow cytometry.
The coimmunoprecipitation (Co-IP) assays were performed as previously described (29). Briefly, total protein was extracted from isolated PMN-MDSCs using immunoprecipitation lysis buffer (Beyotime). Of protein extract, 20% was set apart as input. Co-IP was performed by incubating extract with Pierce Protein G Magnetic Beads (Thermo Fisher Scientific) conjugated with Abs against RBBP5 (catalog no. ab52084; Abcam), CEBPβ (catalog no. ab32358; Abcam), STAT3 (catalog no. 9139; CST), or rabbit IgG (catalog no. ab172730; Abcam) overnight at 4°C. The beads were washed with immunoprecipitation lysis buffer three times and eluted with 100 μl of 1× SDS-PAGE sample loading buffer. After heating at 95°C for 10 min, 20 μl of samples were then subjected to electrophoresis and Western blotting for detection of Rbbp5, Wdr5, Ash2l, Mll1, Cebpβ, p-Stat3, Stat3, or Chop.
Chromatin immunoprecipitation (ChIP) assay was performed using SimpleChIP Enzymatic Chromatin IP Kit (CST) following the manufacturer’s instruction. Abs against H3K4me3 (catalog no. 9751; CST), CEBPβ (catalog no. ab15050; Abcam), and STAT3 (catalog no. 9139; CST) were used for immunoprecipitation. Quantitative analysis of ChIP-derived DNA was performed by semiquantitative PCR and/or real-time quantitative PCR. The assays were performed in triplicate. Primers are as follows: Lamp2-forward: 5′-TCCCCATCAAGGAAAGAGAGT-3′, Lamp2-reverse: 5′-TGGATAGGGAAGCTAGGGGG-3′; miR-21a-forward: 5′-ATTGGTCTATGTCTAGTGGC-3′, miR-21a-reverse: 5′-GTGATGGTGCACGTGTTAAC-3′; miR-21b-forward: 5′-GTCTGTCTGGTCAAAGAAGTC-3′, miR-21b-reverse: 5′-CTGCCTGAGGACCCTAACAG-3′; and miR-181b-forward: 5′-TGATAGTGCAGTGGTCAAGC-3′, miR-181b-reverse: 5′-CTTCCTTTGAGATAGCTTTCC-3′.
All values are presented as the mean ± SEM. Significant differences between the two groups were analyzed by using unpaired Student t test. For multiple groups, significance was evaluated by one-way ANOVA analysis. A p < 0.05 was considered to indicate a statistically significant difference.
Expression of Wdr5, Ash2l, and Mll1 are concurrently downregulated in activated MDSCs both in vitro and in vivo
Initially, the current study established in vitro MDSC induction models by coculturing RBC, B cell–, and T cell–depleted BMCs of BALB/c or C57BL/6 mice with ID8 ovarian carcinoma cells for 2 or 4 d, coculturing BM-MDSCs with 4T1 breast cancer cells for 4 d, or stimulating BM-MDSCs by GM-CSF (40 ng/ml) and IL-6 (40 ng/ml) for 4 d. As shown in Fig. 1A, 1B and Supplemental Fig. 1A, 1B, 2A, 2B, Gr1 and CD11b double–staining was performed to identify murine MDSCs and the results demonstrated that both ID8/4T1–secreted factors and GM-CSF + IL-6 promoted the accumulation of MDSCs. Furthermore, compared with fresh BMCs, activated BM-MDSCs expressed higher mRNA (Fig. 1C) and protein levels (Fig. 1E, Supplemental Fig. 1C) of Arg1 and iNOS, two immunosuppressive mediators produced by expanded MDSCs. Moreover, the essential transcription factors Cebpβ, Chop, and p-Stat3 were also induced by ID8-secreted factors and GM-CSF + IL-6 in BM-MDSCs (Fig. 1D, 1E). These data suggest functional BM-MDSCs were successfully induced by tumor-secreted factors as well as GM-CSF + IL-6. To identify the differentially expressed genes potentially involved in MDSC expansion and activation, the current study applied an Affymetrix microarray to detect the changes of mRNA profiles during MDSC activation induced by ID8-secreted factors for 4 d. Notably, among the differentially expressed genes, the core components of Wdr5-Mll1 complex, including Wdr5, Ash2l, and Mll1 were concurrently decreased in activated BM-MDSCs, compared with untreated BMCs (Fig. 1F), although the expression of another core member of Wdr5/Mll1 complex, Rbbp5, showed no significant change (Fig. 1F). The results were validated by RT-qPCR (Fig. 1G) and Western blot analysis (Fig. 1H, Supplemental Fig. 1D). These findings suggest the Wdr5/Mll1 complex may be destroyed during MDSC activation.
To address whether the observation that Wdr5, Ash2l, and Mll1 downregulation in activated MDSCs in vitro could be extrapolated to MDSCs expanded by tumors in vivo, 5 × 106 B16 melanoma cells were s.c. injected into C57/BL6 mice for 16 d. A significant increase in the proportion of Gr1+CD11b+ MDSCs in the BMCs and splenocytes was observed (Fig. 1I, 1J). As expected, significant downregulation of Wdr5, Ash2l, and Mll1 and elevation of Cebpβ, Chop, and p-Stat3 were observed in the BMCs (Fig. 1K, 1M) and splenocytes (Fig. 1L, 1N) of tumor-bearing mice compared with that in BMCs and splenocytes of normal control mice. To confirm that the change of these genes was indeed attributed to MDSC activation, not to the differentiation of other cells, Gr1+CD11b+ cells were sorted from the BMCs and splenocytes by FACS sorting. The results revealed that the majority of sorted cells were Gr1+CD11b+ cells (>90%) (Fig. 1I, 1J). As expected, similar observation was seen in sorted cells in which both the mRNA and protein levels of Wdr5, Ash2l, and Mll1 were significantly lower in sorted MDSCs derived from BMCs (Fig. 1K, 1M) and splenocytes (Fig. 1L, 1N) of B16-bearing mice than control-sorted MDSCs. Furthermore, the protein levels of Cebpβ, Chop, and p-Stat3 were also significantly higher in sorted BM-MDSCs and spleen MDSCs in B16-bearing mice than control-sorted MDSCs (Fig. 1M, 1N).
To determine whether the changes of WDR5, ASH2L, MLL1, and RBBP5 also occurs in clinical samples. BMCs were isolated from 11 patients with lung cancer and 16 heathy controls. Following depletion of RBCs, CD11b+CD33+HLA-DR− BM-MDSCs were isolated by FACS sorting, and Western blotting were performed to detect the protein expression levels of WDR5, ASH2L, MLL1, and RBBP5. The results demonstrated that the expression levels of WDR5 and ASH2L were also significantly decreased in the BM-MDSCs of patients with lung cancer compared with that of healthy controls (Fig. 2A, 2B). The expression of MLL1 also showed a declining trend, although it was not significant (Fig. 2A, 2B).
Downregulation of Wdr5, Ash2l, and MLL1 contributes to the accumulation and immunosuppressive ability of PMN-MDSCs in vitro
As aforementioned, MDSCs mainly consist of two distinct subsets: CD11b+Ly6ChiLy6G− Mo-MDSCs and CD11b+Ly6ClowLy6G+ PMN-MDSCs. The present study next determined the expression and function of Wdr5, Ash2l, and Mll1 in these two subpopulations of MDSCs. As shown in Fig. 3A, 3B and Supplemental Fig. 2C, 2D, both Mo-MDSCs and PMN-MDSCs were effectively induced in BMCs by GM-CSF + IL-6 as well as ID8-secreted factors. Then, Mo-MDSCs and PMN-MDSCs were sorted in fresh BMCs and GM-CSF + IL-6–treated BMCs. RT-qPCR (Fig. 3C) and Western blotting (Fig. 3D, Supplemental Fig. 2E) results demonstrated the downregulation of Wdr5, Ash2l, and Mll1 only occurred in GM-CSF + IL-6–induced PMN-MDSCs, compared with fresh-sorted PMN-MDSCs.
Afterwards, to characterize the effect of Wdr5, Ash2l, and Mll1 in MDSC induction, the current study constructed lentiviral vectors containing Wdr5, Ash2l, or the C terminus of Mll1 (as the full-length of Mll1 is too large to clone, the C terminus of Mll1 [aa 2623–3967] was constructed, which contains the binding sites for Wdr5, Ash2l, and Mll1 and functional TAD and SET domains). Significantly higher expression levels of Wdr5, Ash2l, and Mll1c were observed postinfection of these lentiviruses in MDSCs (Fig. 3E). Next, the current study determined whether restoration of Wdr5, Ash2l, and Mll1c expression could influence the expansion of MDSCs. As shown in Fig. 3F, 3G and Supplemental Fig. 2A, 2B, the flow cytometry results showed that overexpression of Wdr5, Ash2l, and Mll1c together impaired the induction of MDSC from BMCs by GM-CSF + IL-6 or ID8-secreted factors. Notably, ectopic expression of Wdr5, Ash2l, and Mll1c only could suppress GM-CSF + IL-6 or ID8-secreted factor–induced accumulation of PMN-MDSCs, not Mo-MDSCs (Fig. 3H, 3I, Supplemental Fig. 2C, 2D).
It is known that MDSCs exert its immunosuppressive function via production of several enzymatic mediators, including Arg1 and iNOS. Therefore, the current study determined the expression and activities of Arg and iNOS in PMN-MDSCs isolated from GM-CSF + IL-6–treated BMCs with/without Wdr5 + Ash2l + Mll1c overexpression. As expected, GM-CSF + IL-6 treatment markedly promoted the mRNA (Fig. 3J, 3K) and protein (Fig. 3L) levels of Arg1 and iNOS, as well as the activity of arginase (Fig. 3M) and the production of ROS (Fig. 3N, 3O) in PMN-MDSCs. However, overexpression of Wdr5 + Ash2l + Mll1c in these PMN-MDSCs decreased the expression of Arg1 and iNOS (Fig. 3J–L), decreased the activity of arginase (Fig. 3M) and suppressed the production of ROS (Fig. 3N, 3O).
To gain a more mechanistic insight into the potential role of Wdr5/Ash2l/Mll1c in PMN-MDSCs, CFSE-labeled CD4+ or CD8+ T cells sorted from spleen of healthy mice were cocultured with/without PMN-MDSC cells (2:1 or 4:1 ratio) isolated from GM-CSF/IL-6–treated BMCs with/without Wdr5 + Ash2l + Mll1c overexpression in culture plates precoated with anti-CD3 and CD28 mAbs for 3 d. PMN-MDSCs generated from wild-type GM-CSF + IL-6–treated BMCs efficiently suppressed CD4+ and CD8+ T cell proliferation (Fig. 4A–C). In contrast, PMN-MDSCs generated from GM-CSF + IL-6–treated BMCs with Wdr5 + Ash2l + Mll1c overexpression displayed considerably lower suppressive activity on CD4+ and CD8+ T cell proliferation (Fig. 4A–C). Similar results were observed in ID8-cocultured PMN-MDSCs (Supplemental Fig. 2F, 2G). Additionally, MDSCs can suppress the polarization of M1 macrophages (30–32). In vitro studies were also performed to examine the effect of Wdr5/Ash2l/Mll1c overexpressed PMN-MDSCs on BMDM polarization. BMDMs were cocultured with/without PMN-MDSCs isolated from GM-CSF/IL-6–treated BMCs with/without Wdr5, Ash2l, and Mll1c overexpression for 24 h, then stimulated with LPS and IFN-γ for 24 h. Strikingly, PMN-MDSCs generated from wild-type GM-CSF/IL-6–treated BMCs exert a significantly inhibitory effect on LPS/IFN-γ–induced the expression of M1 macrophage surface marker CD86 and the secretion of IL-6 and IL-12 in BMDMs (Fig. 4D–G). However, PMN-MDSCs isolated from GM-CSF + IL-6–treated BMCs with Wdr5 + Ash2l + Mll1c overexpression showed a markedly decreased ability to suppress LPS + IFN-γ–induced CD86 expression and IL-6 and IL-12 secretion in BMDMs (Fig. 4D–G). Collectively, these data suggest Wdr5 + Ash2l + Mll1c overexpression attenuates the immunosuppressive function of PMN-MDSCs.
Ectopic expression of Wdr5 + Ash2l + Mll1c promotes PMN-MDSCs to differentiate into mature neutrophil-like cells in vitro
Previous studies have demonstrated that PMN-MDSCs are comprised of pathologically activated precursors of neutrophils with vastly different functions (33, 34). Therefore, the current study assessed whether Wdr5 + Ash2l + Mll1c overexpression decreased the immunosuppressive function of PMN-MDSCs via facilitating them to differentiate into mature neutrophil-like cells. One of the classical characteristics of neutrophils is phagocytosis (33). In the current study, the phagocytic activity of PMN-MDSCs isolated from GM-CSF + IL-6–treated or ID8-cocultured BMCs with/without Wdr5 + Ash2l + Mll1c overexpression was assessed with latex beads. Wdr5 + Ash2l + Mll1c overexpression in activated PMN-MDSCs increased their phagocytic activity (Fig. 5A, 5B, Supplemental Fig. 2H, 2I), as well as the expression of phagocytosis marker LAMP2 (Fig. 5C, 5D). To further examine how the MLL1-complex modulates PMN-MDSC maturation, the present study first detected the binding intensity of Wdr5, Ash2l, and MLL1 with Rbbp5 in PMN-MDSCs with/without Wdr5 + Ash2l + Mll1c overexpression. As shown in Fig. 5E and 5F, ectopic expression of Wdr5 + Ash2l + Mll1c restored the binding intensity of Wdr5 and Ash2l with Rbbp5 in PMN-MDSCs. The Co-IP results indicated the Mll1c fragment also interacted with Rbbp5 in activated PMN-MDSCs (Fig. 5F). Furthermore, it was revealed that Wdr5 + Ash2l + Mll1c overexpression enhanced trimethylation of H3K4 in LAMP2 promoter region (Fig. 5G, 5H). These findings suggest that MLL1-complex may promote LAMP2 expression by increasing H3K4me3 in its promoter region. In addition, Wdr5 + Ash2l + Mll1c overexpression also facilitated TNF-α secretion in GM-CSF + IL-6–treated PMN-MDSCs (Fig. 5I, 5J). Taken together, these results support the hypothesis that the MLL1-complex can promote the maturation of PMN-MDSCs, and its disruption is indispensable for the immature status and immunosuppressive function of PMN-MDSCs.
Wdr5 + Ash2l + Mll1c overexpression attenuates the tumor-promoting function of activated PMN-MDSCs in vivo
To determine the function of MLL1-complex in PMN-MDSCs in vivo, a mouse B16 tumor model was established with intratumorous injection of activated PMN-MDSCs with/without Wdr5 + Ash2l + Mll1c overexpression. As shown in Fig. 6A–C, tumors injected with activated PMN-MDSCs with Wdr5 + Ash2l + Mll1c overexpression grew markedly slower than those tumors injected with control activated PMN-MDSCs. Furthermore, FACS analysis of tumor-infiltrating lymphocytes showed that the proportion of both CD4+ and CD8+ T cells were increased in tumors injected with activated PMN-MDSCs overexpressed Wdr5 + Ash2l + Mll1c (Fig. 6D, 6E). In addition, overexpression of Wdr5 + Ash2l + Mll1c in activated PMN-MDSCs also enhanced CD4+IFN-γ+ Th1 and CD8+IFN-γ+ CTL cells in B16 tumor (Fig. 6D, 6E).
GM-CSF + IL-6–induced miR-21a, miR-21b, and miR-181b suppressed the expression of Wdr5, Ash2l, and Mll1 by targeting to their 3′-UTRs, respectively, and modulates the accumulation and function of PMN-MDSCs
Multiple studies over the past decades have revealed the fundamental role of miRNAs in the differentiation and function of MDSCs (35, 36). To define regulators that could modulate the expression of Wdr5, Ash2l, and Mll1 in MDSCs, the current study looked through the miRDB Database (http://mirdb.org/) and identified miRNAs with putative binding sites in the 3′-UTRs of Wdr5, Ash2l, and Mll1. Notably, it was revealed that miR-21 and miR-181b, which have been reported to be synergistically induced by STAT3-CEBPβ–transcriptional complex in sepsis-associated MDSCs (23, 25), have potential binding sites within the 3′-UTRs of Wdr5, Ash2l, and Mll1, respectively (Fig. 7A). To determine whether the 3′-UTRs of Wdr5, Ash2l, and Mll1 are targeted by miR-21a, miR-21b, or miR-181b directly, the 3′-UTR fragments of Wdr5, Ash2l, and Mll1 containing the miR-21a, miR-21b, or miR-181b binding site(s) were cloned into the psiCHECK2 dual luciferase reporter plasmid and were transfected into 293T cells along with control miRNA, miR-21a, miR-21b, or miR-181b mimics (Fig. 7B). The results of the luciferase assay demonstrated that enhanced expression of miR-21a, miR-21b, or miR-181b dramatically decreased the luciferase activity of Wdr5, Ash2l, or Mll1 3′-UTR wild-type constructs, respectively (Fig. 7C–E). Afterwards, the wild-type constructs were mutated as depicted in Fig. 7A and also transfected into 293T cells along with miRNA mimic, miR-21a, miR-21b, or miR-181b mimics. The results showed that mutation of the binding sties abolished the effects of miR-21a, miR-21b, or miR-181b on the luciferase activity of Wdr5, Ash2l, or Mll1 3′-UTR plasmids, respectively (Fig. 7C–E). Notably, the results revealed that ectopic expression of miR-21a, miR-21b, and miR-181b with transfection of corresponding miRNA mimics (Fig. 7F–H) significantly inhibited the protein levels of Wdr5, Ash2l, and Mll1 in isolated PMN-MDSCs from BMCs, respectively (Fig. 7I–K). To determine the transfection efficiency and cytotoxicity of miRNA mimics in PMN-MDSC, we transfected with freshly isolated PMN-MDSCs with 0, 20, 50, and 100 nM of FAM-labeled miRNA mimics control. The results revealed that miRNA mimic was successfully transfected (Supplemental Fig. 3A, 3B) into PMN-MDSCs with low cytotoxicity (Supplemental Fig. 3C).
To address whether the downregulation of Wdr5, Ash2l, and Mll1 could be, at least partially, attributed to upregulation of miR-21a, miR-21b, and miR-181b in activated BM-MDSCs and PMN-MDSCs, the current study then detected the expression of these miRNAs in BM-MDSCs with/without GM-CSF + IL-6 treatment. As expected, all the expression levels of miR-21a, miR-21b, and miR-181b were upregulated in GM-CSF + IL-6–induced BM-MDSCs compared with fresh BMCs (Fig. 8A). Inhibition of these miRNAs (Fig. 8A) restored the expression of Wdr5, Ash2l, and Mll1, respectively, in GM-CSF + IL-6–treated BM-MDSCs (Fig. 8B). Functionally, suppression of either miR-21a, miR-21b, or miR-181b attenuated GM-CSF + IL-6–induced BM-MDSC expansion (Fig. 8C, 8D) as well as the proportion of PMN-MDSCs (Fig. 8E, 8F) in activated BMCs. In vivo tumor experiment also revealed that tumors injected with GM-CSF + IL-6–treated PMN-MDSCs grew significantly faster than those injected with fresh PMN-MDSCs (Fig. 8G–I). However, inhibition of either miR-21a, miR-21b, or miR-181b in activated PMN-MDSCs could slow down the tumor growth rate (Fig. 8G–I). Taken together, the findings of Figs. 7, 8 suggest that miR-21a, miR-21b, and miR-181b may facilitate PMN-MDSC expansion and activation via downregulating Wdr5, Ash2l, and Mll1, respectively.
p-Stat3 and Cebpβ synergistically promote the transcription of miR-21a, miR-21b, and miR-181b and suppress the expression of Wdr5, Ash2l, and Mll1 in activated PMN-MDSCs
It has been reported that Stat3 and CEBPβ coordinatively induce miR-21 and miR-181b expression in sepsis MDSCs (25). Therefore, the current study then assessed whether p-Stat3 and Cebpβ synergistically suppressed the expression of Wdr5, Ash2l, and Mll1 by inducing the transcription of miR-21a, miR-21b, and miR-181b, respectively. The Western blotting results indicated that silence of Cebpβ enhanced the protein levels of Wdr5, Ash2l, and Mll1, whereas silence of Stat3 only enhanced the protein levels of Wdr5 in activated PMN-MDSCs (Fig. 9A). Notably, when both Stat3 and Cebpβ were knocked down, both the mRNA and protein levels of Wdr5, Ash2l, and Mll1 were significantly increased in activated PMN-MDSCs (Fig. 9B, 9C). Furthermore, Co-IP experiments proved that Cebpβ physically interacted with p-STAT3 (Tyr705) in activated PMN-MDSCs (Fig. 9D, 9E). The results also confirmed that silencing of Stat3 and/or Cebpβ significantly suppressed the expression of miR-21a, miR-21b, and miR-181b in activated PMN-MDSCs (Fig. 9F). The present study also performed a ChIP assay to determine the binding potential of p-Stat3 and Cebpβ on the promoter regions of miR-21a, miR-21b, and miR-181b. In PMN-MDSCs isolated from fresh BMCs, no binding bands of p-Stat3 or Cebpβ could be detected. Nevertheless, both p-Stat3 and Cebpβ showed strong binding activities on the promoter regions of these miRNAs in activated PMN-MDSCs (Fig. 9G).
In conclusion, the results of the current study reveal a critical regulatory axis in PMN-MDSC expansion, activation, and maturation, in which tumor environment–associated factors such as GM-CSF and IL-6 activate transcription factors STAT3 and CEBPβ, and thus enhance the transcription of miR-21a, miR-21b, and miR-181b. Increased levels of these miRNAs suppressed the expression of Wdr5, Ash2l, and Mll1 to disrupt the Mll1-complex. Disruption of MLL1-complex is beneficial to PMN-MDSCs to maintain its immature status and immunosuppressive function (Fig. 10). This axis may provide an effective immunological therapeutic approach for patients with tumor or other immune-associated diseases.
MDSCs are a heterogeneous population of myeloid cells with immature and undifferentiated phenotype. Accumulating evidence has revealed that MDSCs exert its immunosuppressive ability in numerous pathological conditions, such as different types of cancer and inflammatory diseases. Inhibition of MDSC function may be an effective therapeutic strategy for these diseases (37–39). However, the intrinsic mechanisms underlying MDSC expansion, differentiation, and activation urgently requires further investigation. The present study revealed that the core members of the Mll1-complex, Wdr5, Ash2l, and Mll1 were concurrently suppressed by GM-CSF and IL-6 in PMN-MDSCs. Downregulation of Wdr5, Ash2l, and Mll1 impaired the formation of the Mll1-complex and is indispensable for maintaining the immature phenotype and immunosuppressive function of activated PMN-MDSCs. Furthermore, p-Stat3– and CEBPβ-induced miR-21a, miR-21b, and miR-181b play an inhibitory role in the expression of Wdr5, Ash2l, and Mll1, respectively. Taken together, the results of the current study have demonstrated a key role of Stat3/Cebpβ–miR-21a/b/181b–Mll1-complex in regulating the differentiation and function of PMN-MDSCs.
The results of protein expression of WDR5, ASH2L, MLL1, and RBBP5 in BM-MDSCs of patients with lung cancer and that of healthy controls showed that the expression of WDR5 and ASH2L were also significantly decreased in BM-MDSCs of patients with lung cancer compared with that of healthy controls. The expression of MLL1 also showed a declining trend, although it was not significant, which suggest more clinical samples should be used.
Over the past few decades, the mechanisms by which the MLL1-complex modulates leukemogenesis have been deeply investigated. However, its function in MDSCs remains unknown. To the best of our knowledge, these data are the first to demonstrate the critical role of MLL1-complex in MDSCs. As a pivotal member of the histone methyltransferases family, MLL1 encodes an enzyme catalyzing the mono-, di-, and trimethylation of H3K4 to control gene transcription. Alteration of this process often causes changes in the gene expression patterns involved in stem cell maintenance and development. The present study hypothesized that disruption of the Mll1-complex may influence the genes associated with PMN-MDSC differentiation and function. In fact, the current study indeed proved that the Mll1-complex enhanced the transcription and expression of phagocytosis-associated genes LAMP2 by enhancing H3K4me3 in its promoter region, and thus facilitated PMN-MDSC maturation. In contrast, several critical cytokines and transcription factors involved in MDSC expansion and activation have been reported to be regulated by MLL1 in other immune cells. For instance, deletion of MLL1 in T cells and macrophages impairs the production of lineage-specific cytokines, such as TNF-α, IFN-γ, IL-1β, and IL-4, concurrently with the dysregulated transcriptional activities of NF-κB and GATA3 (40, 41). In the current study, it was revealed that overexpression of Wdr5 + Ash2l + Mll1C in murine PMN-MDSCs suppressed Arg1 and iNOS and elevated the secretion of TNF-α. However, H3K4me3 was not detectable in their promoter regions, which suggests that these genes may be regulated by Mll1-complex indirectly. Future studies will address this problem. In contrast, other genes and cytokines regulated by the MLL1-complex in PMN-MDSCs also require investigation. Future studies will be focused on the epigenetic events controlled by MLL1-complex on the whole genome during PMN-MDSC accumulation and activation, particularly on the histone methylation profiles.
One of the interesting findings in the current study is that we found that overexpression of Wdr5 + Ash2l + Mll1C might promote PMN-MDSCs differentiating into mature neutrophil-like cells. In our future study, it is interesting to determine whether these mature neutrophil-like cells exert their phagocytic ability to erase tumor cells directly or present tumor-associated Ags to T cells in tumor microenvironment to activate tumor immune response.
Another interesting finding in the current study is that the components of Mll1-complex, Wdr5, Mll1, and Ash2l are concurrently downregulated in MDSC expansion and activation. The mechanism underlying this phenomenon is worth investigating. Numerous studies have shown that a wide range of miRNAs implicate in MDSC generation and differentiation. For example, as a downstream target of Stat3 and Cebpβ, miR-181b couples with miR-21 to generate MDSCs and enhances its immunosuppressive ability in early and late sepsis (23, 25). Notably, by using online software miRDB, the current study revealed that the 3′-UTRs of Wdr5, Ash2l, and Mll1 have potential binding sites for miR-21a, miR-21b, and miR-181b, respectively. The present study also demonstrated that suppression of miR-21a, miR-21b, and miR-181b could reverse GM-CSF + IL-6–induced PMN-MDSC accumulation, similar to the effect of Wdr5 + Ash2l + Mll1c overexpression. Meanwhile, luciferase assay proved that miR-21a, miR-21b, and miR-181b directedly targeted to the 3′-UTRs of Wdr5, Ash2l, and Mll1, respectively. Furthermore, overexpression of miR-21a, miR-21b, and miR-181b indeed inhibited the expression of Wdr5, Ash2l, and Mll1, respectively. Even so, other factors involved in regulating Wdr5, Ash2l, and Mll1 expression in MDSCs also need to be identified.
There are several limitations to the current study. As aforementioned, first, the downstream effectors of Mll1-complex during MDSC expansion and activation should be further investigated. Second, the upstream regulators of Wdr5, Mll1, and Ash2l need to be further identified. Third, the dynamic changes of interactions among Wdr5, Mll1, Ash2l, and other transcriptional factors should be deeply investigated during the differentiation and maturation of PMN-MDSCs. In addition, clinical studies for more specimens are lacking for the study. Future studies will focus more on the role of the Mll1-complex in PMN-MDSCs of cancer patients.
In conclusion, the results of the current study reveal the suppressive role of Mll1-complex in murine MDSCs and its upstream regulators, as well as downstream effectors. It is believed that the work described in the current study not only provides a novel role for MLL1-dependent H3K4me3 chromatin remodeling in immunosuppressive ability of MDSCs, but also demonstrates an effective therapeutic target for inflammatory diseases and cancer.
This work was supported partially by China Postdoctoral Science Foundation Grant 2015M581292 and National Natural Science Foundation of China Grant 31470876 and was self-financed partially by Z.Z.
The online version of this article contains supplemental material.
Abbreviations used in this article:
bone marrow cell
bone marrow–derived macrophage
bone marrow MDSC
Cell Signaling Technology
histone H3 at lysine 4
NO synthase 2
myeloid-derived suppressor cell
mixed-lineage leukemia 1
reactive oxygen species
small interfering RNA
3′ untranslated region.
The authors have no financial conflicts of interest.