Abstract
Long noncoding RNA (lncRNA) plays a critical role in many biological processes, such as cell differentiation and development. However, few studies about lncRNAs regulating the differentiation and development of myeloid-derived suppressor cells (MDSCs) exist. In this study, we identified a lncRNA pseudogene, Olfr29-ps1, which was expressed in MDSCs and upregulated by the proinflammatory cytokine IL6. The Olfr29-ps1 in vertebrates is conserved, and the similarity between the Olfr29-ps1 and human OR1F2P sequence is 43%. This lncRNA promoted the immunosuppressive function and differentiation of monocytic (Mo-)MDSCs in vitro and in vivo. It directly sponged miR-214-3p to downregulate miR-214-3p, which may target MyD88 to modulate the differentiation and development of MDSCs. The functions of Olfr29-ps1 were dependent on IL6-mediated N6-methyladenosine (m6A) modification, which not only enhanced Olfr29-ps1, but also promoted the interaction of Olfr29-ps1 with miR-214-3p. Thus, our results demonstrated that the pseudogene Olfr29-ps1 may regulate the differentiation and function of MDSCs through a m6A-modified Olfr29-ps1/miR-214-3p/MyD88 regulatory network, revealing a mechanism for the regulation of myeloid cells and also providing potential targets for antitumor immunotherapy.
Introduction
Myeloid-derived suppressor cells (MDSCs) are a major regulator of immune responses in cancer and inflammation. These MDSCs, which are derived from bone marrow progenitor cells, are a group of heterogeneous cells with immunosuppressive functions, including immature granulocytes, dendritic cells, macrophages, and early undifferentiated myeloid precursor cells. Tumors, inflammation, or infection may result in the accumulation and expansion of MDSCs (1–3), which are induced by diverse cytokines such as GM-CSF, G-CSF, VEGF, and IL6 (4). These cells are identified as CD11b+Gr1+ cells, which are further divided into two subsets, including polymorphonuclear MDSCs (PMN-MDSC) identified as CD11b+Ly6G+Ly6Clo cells and monocytic MDSCs (Mo-MDSC) identified as CD11b+Ly6G−Ly6Chi cells (5). Human MDSCs, including Mo-MDSCs (CD14+) and PMN-MDSCs (CD15+), are described as lineage-negative cells that coexpress CD11b and CD33 but lack HLA-DR. MDSCs inhibit the immune response of T cells and mediate immunosuppression by the expression of arginase-1 (Arg-1), NADPH oxidase 2 (NOX2), nitric oxide synthase 2 (NOS2), COX2, and production of nitric oxide (NO) and reactive oxygen species (ROS; ref. 6).
Long noncoding RNAs (lncRNA) have an important role in diverse biological processes by regulating gene expression in cis or in trans (7–9). They could be from a wide variety of transcripts, including intergenic and intragene transcripts, natural antisense chains, various enhancers, and promoter transcripts (8, 10, 11). Studies show that lncRNAs can regulate the development and differentiation of immune cells by a variety of mechanisms (12–17). We have found that the lncRNA HOTAIRM1 can modulate peripheral blood cells to differentiate into dendritic cells (DC) by sponging miR-3960 to regulate HOXA1 expression (18), and have demonstrated that IL6-mediated RNCR3 and lnc-chop may affect MDSC development (19, 20). Studies have also found that some pseudogene transcripts can function as lncRNAs to regulate related gene expression by different mechanisms (13, 21–23). We, here, demonstrated that the pseudogene Olfr29-ps1 may promote the immunosuppressive function and differentiation of Mo-MDSCs by sponging miR-214-3p after N6-methyladenosine (m6A) modification.
Materials and Methods
Mice, human samples, and cell lines
C57BL/6 mice were purchased from the Beijing Animal Center (Beijing, China) and maintained in a specific pathogen-free facility. B6.129S6-Il-6tm1Kopf (IL6−/−) and B6.SJL-CD45a(Ly5a; CD45.1) mice were purchased from the Model Animal Research Center of Nanjing University (Nanjing, Jiangsu, China). OT-I or OT-II mice were from Dr. Linrong Lu (Zhejiang University). All animal experiments were carried out in accordance with the Nankai University Guide for the Care and Use of Laboratory Animals and with the approval of the Nankai University Animal Care and Use Committee.
The peripheral blood and tissue samples from patients with colon or rectal adenocarcinoma, which were demonstrated according to pathologic criteria, were obtained after informed consent at People Union Hospital (Tianjin, China). Same age and sex healthy human peripheral blood control samples were obtained after signing informed consents. The collection and use of all human samples (healthy individuals and patients with colon or rectal adenocarcinoma) were approved by the Institute's Human Ethics Committee of Nankai University in accordance with the Declaration of Helsinki. The samples from 12 patients with colon cancer, 8 patients with rectal cancer, and 8 healthy individuals (45–60 years old, male) were immediately analyzed for peripheral blood cells and tumor tissues were stored at −70°C (not for more than 1 month). These patients met the following criteria: colon cancer or rectal cancer was determined by pathologic examination and that the patients did not receive drug treatment inside 1 month.
The murine melanoma B16, human monocyte cell line U937, and human embryonic kidney cell line HEK 293T cells were obtained from the American Type Culture Collection during 2013 to 2014. The murine ovarian tumor cell line 1D8 was from Dr. Richard Roden (The Johns Hopkins University School of Medicine, gift in 2010). These cell lines were authenticated by the short-tandem repeat method but not further authenticated in the past years. They were not contaminated by Mycoplasma before or after experiments. These cells were cultured in RPMI-1640 medium with 10% FCS and 1% penicillin and streptomycin (P/S).
Plasmids, siRNA, and microRNA transfection
A total of 1 × 107 C57BL/6 bone marrow cells (BMC) were collected from femur and then were transfected with pcDNA3.1 (5 μg/mL), pcDNA3.1-METTL3 (5 μg/mL), control scrambled siRNA (100 nmol/L), Olfr29-ps1 siRNA (100 nmol/L), MyD88 siRNA (100 nmol/L), METTL3 siRNA (100 nmol/L), miR-214-3p mimics (100 nmol/L), miR-214-3p inhibitor (100 nmol/L, chemically modified small RNA for cell-specific target microRNA), miR-761 mimics (100 nmol/L, chemically synthesized miRNA sequence) or scrambled control using HiPerFect transfection reagent (Qiagen) according to the manufacturer's instructions, and then cultured in RPMI-1640 medium with 10% FCS and 1% P/S for 4 days in the presence of GM-CSF (40 ng/L) plus IL6 (40 ng/L). All microRNAs, siRNAs, and control siRNAs were purchased from Riobio. PcDNA3.1-METTL3 was generated by cloning METTL3 (ID: 56335) and conjugating into pcDNA3.1/V5-His TOPO TA vector (Invitrogen). The target sequences for Olfr29-ps1 siRNA, MyD88 siRNA and METTL3 siRNA, as well as sources, are listed in Supplementary Table S1. For constructing recombinant gene-expression plasmids, the full-length sequence of Olfr29-ps1 (ID: 29848) or OR1F2P (ID: 26184) was amplified using PCR methods (primer pairs are described in Supplementary Table S1). The PCR products were directly cloned into the pcDNA3.1/V5-His TOPO TA plasmid (Invitrogen) using T4-conjugating enzyme (BioMart), which was named M-Olfr29-ps1 or Hu-OR1F2P.
For JAK1 and STAT3 inhibitor–treated MDSCs, MDSCs, which were induced according to the described protocol in this method, were incubated with the JAK1 inhibitor filgotinib (20 nmol/L; Selleckchem) or the STAT3 inhibitor HO-3867 (100 nmol/L; Selleckchem) for 24 hours.
Lentivirus construction and transduction
A short hairpin RNA (shRNA) target sequence (5′-GCTGTCTCTGTGGTTCAAA-3′) of Olfr29-ps1 was chosen by BLOCK-iT RNAi Designer (Invitrogen) and/or by i-Score Designer38 (https://www.med.nagoya-u.ac.jp/neurogenetics/i_Score/i_score.html). The Olfr29-ps1 shRNA constructs were made using pGreenPuro shRNA cloning and expression lentivector kit (System Biosciences Inc.) according to the manual. The control shNC was a luciferase control shRNA from the kit. For packaging lentivirus particles, the shRNA lentivector or Olfr29-ps1 lentivector together with pMD2.G and psPAX2 packaging plasmids (Invitrogen) were cotransfected into 293T cells. MDSCs (1 × 107) were infected with the lentiviral supernatants in the presence of polybrene (8 μg/mL; Millipore) by centrifugation and then cultured with RPMI-1640 medium with 10% FCS and 1% P/S for 24 hours. The cells were then washed and cultured in the presence of GM-CSF (40 ng/mL) plus IL6 (40 ng/mL) for 4 days.
In vitro induction of MDSCs and macrophages
For in vitro induction of MDSCs, BMCs were obtained from the femurs of C57BL/6 mice and cultured in RPMI-1640 medium supplemented with GM-CSF (40 ng/mL) only or GM-CSF (40 ng/mL) plus IL6 (40 ng/mL) for 4 days. We also induced MDSCs in vitro as above, and then we harvested the cells at 0 hour, 12 hours, 24 hours, and 48 hours, respectively. For the detection of the IL6 dose-dependence of Olfr29-ps1, we set up different concentrations (including 10, 20, 40, and 60 ng/mL) of IL6 in combination with GM-CSF (40 ng/mL) stimulation to induce MDSC production. To prepare tumor cell supernatant-induced CD11b+Gr1+ MDSCs in vitro, 5 × 104 1D8 or B16 tumor cells (upper chamber) were cocultured with 2 × 106 BMCs (lower chamber) in a 24-transwell plate in the presence of GM-CSF (40 ng/mL) for 4 days. CD11b+Gr1+cells were sorted by FAScan or isolated using CD11b and Gr1 MACS MicroBeads and cell isolation kit (Miltenyi Biotec) according to the manufacturer's instructions. Macrophages were induced in vitro under the presence of M-CSF (40 ng/mL) for 4 days.
Human MDSC-like cells were generated through human peripheral blood monocytes cultured with GM-CSF and IL6 according to previously reported methods (24). Human peripheral blood mononuclear cells were isolated using a CD14+ isolation kit (R&D Systems), and then cultured in RPMI-1640 medium with 10% FCS and 1% P/S in the presence of human recombinant GM-CSF (40 ng/mL) and IL6 (40 ng/mL) for 4 days.
MDSCs were from C57BL/6 or C57BL/6 IL6 KO mice. B16 tumor cells were injected into wild-type (WT) and IL6 KO mice. PBS was used as a control. After 4 weeks, MDSCs were sorted through staining CD11b and Gr1 from WT and IL6 KO mice.
M-MDSCs and PMN-MDSCs were sorted using flow cytometry through staining CD11b and Ly6C or Ly6G from Lv-shOlfr29-ps1–transduced MDSCs or Lv-oeOlfr29-ps1–transduced MDSCs. The tumors were subcutaneously dissected from the groin of mice.
Bioinformatic analyses
All miRNAs were downloaded from miRbase (miRBase V.20, www.mirbase.org). We predicted the interactions of miRNAs with Olfr29-ps1 by two computational algorithms: RNAHybrid program (https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid) and miRDB (http://mirdb.org/index.html).
Flow cytometry
MDSCs (1 × 106) from in vitro culture or BMCs of mice with tumors were collected and washed twice with PBS, and then incubated in PBS with 1% FBS blocking and antibodies for 30 minutes. After washing twice with PBS, the cells were fixed in 1% paraformaldehyde and analyzed by FACScan flow cytometer (BD Biosciences). Dead cells were eliminated through 7-AAD (BD Biosciences) staining. All antibodies used in this study were listed in Supplementary Table S1.
Real-time qPCR and PCR
Total RNA was extracted from MDSCs by using TRIzol reagent (Life Technologies) and was transcribed to cDNA using the HiFiScript cDNA Synthesis Kit (CWBIO) according to the manufacturer's instructions. The quantitative real-time PCR (qRT-PCR) was performed by using HieffqPCR SYBR-Green Master Mix (YEASEN) in a Bio-Rad iQ5 multicolor RT-PCR system. The primers used for qRT-PCR were shown in Supplementary Table S1. The expression of each gene, including those encoding MyD88, OR1F2P, Arg-1, iNOS, lncRNA Olfr29-ps1, and miRNAs miR-214-3p, miR-149-5p, and miR-361-3p, was calculated using the 2−ΔΔCT method. The stem-loop RT primer (BGI) was used for miR-214-3p or other miRNA reverse transcription. The relative expression of miRNAs was normalized to that of the internal control U6. For other genes, GAPDH was used as the endogenous control. RNA content in the sample was detected using NanoDrop. One hundred nanograms/reaction was used with three replicates. The full-length sequence of Olfr29-ps1 was amplified using an RT-PCR amplification kit (Takara; primer pairs are described in Supplementary Table S1).
Western blot
Western blot was performed as described previously (25). Briefly, cells were harvested at the indicated times and rinsed twice with ice-cold PBS. The cells were lysed with cell-lysis buffer (Cell Signaling Technology) and centrifuged at 14,000 × g for 10 minutes at 4°C. The protein concentrations of the extracts were measured using a bicinchoninic acid assay (Pierce). Thirty micrograms of protein was loaded into gels and then wet transferred. Hybridizations with 1 μg of primary antibodies (Abs) were carried out for 1 hour at room temperature in blocking buffer (TBS with 5% skim milk powder). Antibodies against iNOS (Cell Signaling Technology; 1:1,000 dilution), Arg-1 (Santa Cruz Biotechnology; 1:1,000 dilution), Nox2 (Santa Cruz Biotechnology; 1:1,000 dilution), Cox2 (Cell Signaling Technology; 1:2,000 dilution), MyD88 (Santa Cruz Biotechnology; 1:1,000 dilution), and β-actin (Santa Cruz Biotechnology; 1:1,000 dilution) were used. The protein–Ab complexes were detected using peroxidase-conjugated secondary Abs (1:5,000 dilution; Boehringer Mannheim) and enhanced chemiluminescence (ECL+; Amersham Biosciences). The signals were checked by autoradiography film when the ECL substrate was added to the membranes. The primary and secondary antibodies used in this study were listed in Supplementary Table S1.
Arginase activity, nitric oxide, H2O2, and ROS detection
For arginase activity, MDSCs induced from BMCs of C57BL/6 mice (5 × 106) were lysed for 30 minutes with 100 mL of 0.1% Triton X-100 (Sigma-Aldrich) at 4°C. Following lysis, 100 μL of Tris-HCl (25 mmol/L) and 10 mL of MnCl2 (10 mmol/L) were added, and the mixture was heated for 10 minutes at 56°C. Subsequently, the 100 μL lysates were incubated with 100 mL of 0.5 mol/L L-arginine (Sigma-Aldrich) at 37°C for 120 minutes. The reaction was stopped with 900 mL of H2SO4 (96%)/H3PO4 (85%)/H2O (1:3:7). Urea concentration was measured by absorbance at 540 nm (Full-Wavelength Enzyme Marker, Multiskan Sky) after the addition of 40 mL of 9% α-isonitrosopropiophe, followed by heating at 95°C for 30 minutes. A standard curve was generated using serial dilutions of 120 mg/mL urea (120, 12, 1.2, and 0.12 mg/mL urea). Arginase activity (unit) was defined by the amount of enzyme that catalyzes the formation of 1 mg of urea per minute.
For nitric oxide production, the total nitric oxide in the 60 μL cell lysates was measured using the Nitrate/Nitrite Assay Kit (Kamiya). Equal volumes of cell lysates (60 mL), NADPH (2 mmol/L, 5 mL, Beyotime), FAD (10 mL, Beyotime), and nitrate reductase (5 mL, Beyotime) were incubated at 37°C for 30 minutes, followed by the addition of 10 mL of LDH buffer (Abcom). After incubation for 30 minutes at 37°C, 50 mL of Griess Reagent I (Beyotime) and Griess Reagent II (Beyotime) was added, incubated at room temperature for 10 minutes, and measured at 540 nm (Full-Wavelength Enzyme Marker, Multiskan Sky). Nitrite concentrations were quantified by comparing the absorbance values with a standard curve generated by serial dilutions of 1 mol/L sodium nitrite (2, 5, 10, 20, 40, 60, and 80 μmol/L).
For H2O2 production, H2O2 was evaluated using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen). Briefly, 1 × 104 MDSCs induced in vitro from BMCs of C57BL/6 mice were resuspended in Krebs-Ringer phosphate (50 mmol/L Amplex Red reagent and 0.1 U/mL HRP). After the addition of PMA (phorbolmyristate acetate, 30 ng/mL), the absorbance at 560 nm was measured using a microplate reader at 37°C (Full-Wavelength Enzyme Marker). Absorbance values for the test samples were normalized to a standard curve generated by serial dilutions of 10 mmol/L H2O2 (0, 12.5, 25, 50, or 75 μL of 5 mmol/L H2O2 in 0.5 mL buffer).
For ROS detection, ROS production by MDSCs was measured by using oxidation-sensitive dye DCFDA (diacetyldichlorofluorescein, Molecular Probes/Invitrogen) according to the reported protocol (26). MDSCs (1 × 106) were incubated at 37°C in RPMI medium in the presence of DCFDA (2.5 mmol/L) for 30 minutes. For PMA-induced activation, cells were simultaneously cultured with DCFDA and PMA (30 ng/mL), and ROS expression was analyzed by staining DCFDA (antibodies listed in Supplementary Table S1).
In vitro MDSC immunosuppressive function
To measure the immunosuppressive function of MDSCs transduced with Olfr29-ps1 shRNA or Olfr29-ps1/lentiviruses, the splenocytes obtained from OT-I or OT-II mice were cocultured with MDSCs in the presence of 200 nmol/L OVA peptide (OVA257-264, GenScript) or OVA peptide (OVA323-339, GenScript) in 96-well plates at a ratio of 1:0, 1:1, 1:1/2, 1:1/4, 1:1/8, and 0:1 for 48 hours. The production of IFNγ was measured by an ELISA Kit (Biotech) according to the manufacturer's instructions. For ELISA, 100 μL undiluted supernatants were used.
In vivo experiments
The C57BL/6 B16 melanoma mouse model was used to investigate the effect(s) of Olfr29-ps1–modified MDSCs on tumor growth. Mice were injected with 1 × 106 B16 cells via subcutaneous injection at an inguinal site and were randomly divided into Olfr29-ps1 shRNA/lentivirus, Olfr29-ps1/lentivirus (Lv-oeOlfr29-ps1), and overexpressing and knockdown control lentivirus groups (6 mice/group), and then the isolated MDSCs (1 × 106) were injected into different groups via the tail vein after injection of tumor cells. For the preparation of Olfr29-ps1–modified MDSCs, 1 × 107 BMCs obtained from C57BL/6 CD45.1 mice were transduced with Olfr29-ps1 shRNA or Olfr29-ps1/lentiviruses [1.5 × 108 transducing units (UT)/mL], and then cultured with GM-CSF (40 ng/mL) plus IL6 (40 ng/mL) for 4 days. The tumor volume was measured in two dimensions by calipers every 2 days and calculated by the following formula: Width2 × Length × π/6. Twenty-four days later, the tumors were dissected from the groin of the tumor mice. After grinding, CD4 + T cells and CD8 + T cells, CD11b+Gr1+ cells, and CD11b+Ly6G+Ly6C+cells were analyzed by flow cytometry.
RNA immunoprecipitation
RNA immunoprecipitation (RIP) was performed according to a previously reported protocol (27). Briefly, the GM-CSF plus IL6-induced MDSCs were harvested and washed. Ice-cold IP lysis buffer (Thermo Scientific Pierce) containing 0.5% ribonuclease inhibitor (Invitrogen) was then added, and the cells were incubated on ice for 5 minutes with periodic mixing. The lysates were then transferred into a microcentrifuge tube and centrifuged at 13,000 × g for 10 minutes at 4°C to pellet cell debris, and the supernatants were transferred into a new tube, and protein G agarose (Supplementary Table S1) was added and incubated for 1 hour at 4°C with rotation for preclearing. The anti–N6-methyladenosine (m6A; 1 μg, Abcam) and anti-argonaute-2 antibody mouse/human (1 μg, Abcam) were added and incubated overnight at 4°C with rotation. Protein G agarose was pelleted by brief centrifugation (3,000 × g for 1 minute) and then washed sequentially with IP lysis buffer (containing 0.5% ribonuclease inhibitor). Finally, RNA was extracted from protein/RNA complexes bound with the beads using TRIzol reagent and dissolved in DEPC water and quantified by quantitative PCR (qPCR) as described above. The RNA IP PCR-specific primers are listed in Supplementary Table S1.
RNA–protein pulldown analyses
RNA–protein pulldown analyses were performed using the Pierce Magnetic RNA–protein pulldown Kit. MDSCs induced in vitro from BMCs of C57BL/6 mice were harvested, and cell lysates were prepared using IP lysis buffers (Thermo Scientific Pierce) according to the manufacturer's protocol. Olfr29-ps1 was transcribed (NEB, manual HiScribe T7 in vitro transcription Kit) and labeled using the RNA Desthiobiotinylation Kit (Thermo Scientific Pierce) in vitro. Fifty microliters of beads and 50 pmol/L of labeled RNA were added into RNA capture buffer and incubated for 30 minutes at room temperature with agitation to bind labeled Olfr29-ps1 to streptavidin magnetic beads. After washing beads with an equal volume of Tris (20 mmol/L, pH 7.5), 100 mL of 1× protein–RNA binding buffer was added into the beads and mixed. The master mix (100 mL) of the RNA–protein binding reaction was added to the RNA-bound beads, mixed by pipetting, and then incubated 60 minutes at 4°C with rotation to bind the proteins to RNA. After washing the beads twice with 100 mL of wash buffer, 50 mL of elution buffer was added and incubated 30 minutes at 37°C with agitation. The samples obtained were used for immunoblotting. Ago2 was used for the primary body.
Immunostaining and RNA fluorescence in situ hybridization
Immunostaining and RNA fluorescence in situ hybridization (RNA-FISH) was performed according to the reported protocol (20). MDSCs or human peripheral blood monocyte cells were first slicked on sterile and 0.01% polylysine–treated slides in the bottom of a 6-well tissue culture dish. The slides were then processed sequentially with ice-cold CSK buffer (cytoskeletal (CSK) buffer containing 100 mmol/L NaCl, 300 mmol/L sucrose, 3 mmol/L MgCl2, and 10 mmol/L PIPES pH 6.8 at room temperature, 20–25°C), 0.4% Triton X-100 buffer, and CSK buffer for 30 seconds for cell membrane perforation. The slides were then treated with 4% PFA for 10 minutes and cold 70% ethanol three times for cell fixation. After washing three times with ice-cold PBS, the slides were blocked in prewarmed 5% goat serum (Abcom) for 30 minutes at 37°C, and the slides were then incubated with CD11b at 37°C for 1 hour, washed three times with 1X PBS/0.2% Tween-20 for 3 minutes on a rocker, and then incubated with goat anti-Mouse IgG H&L (Abcom) at 37°C for 30 minutes. After washing three times with 1X PBS/0.2% Tween-20, the slides were fixed with 2% PFA at room temperature for 10 minutes. The slides were dehydrated by moving them through a room temperature ethanol series (85%, 95%, and 100% ethanol) for 2 minutes each, and air dried at room temperature for 15 minutes. The slides were then hybridized using the indicated probes overnight at 37°C in a humid chamber. After washing with 2× SSC/50% formamide, 2 × SSC, and 1 × SSC, each for three times, DAPI dye was added. Finally, the slides were sealed and then observed using a confocal microscope (Olypus FV1000).
Dual-luciferase assay
Luc-Olfr29-ps1 plasmids were constructed by cloning the sequence of Olfr29-ps1 into the downstream of a firefly luciferase cassette in the pSiCHECK-2 vector (Promega). The primers used were listed in Supplementary Table S1. HEK293T cells and MDSCs were, respectively, cultured in a 24-well plate at 1 × 105 cells per well. HEK293T cells (1 × 107) were cotransfected with Luc-Olfr29-ps1 (1 μg/mL) and 100 nmol/L miR-214-3p mimic or mimic control; MDSCs were cotransfected with luc-Olfr29-ps1 (1 μg/mL) and 100 nmol/L miR-214-3p inhibitor or inhibitor control by using Lipofectamine 2000 (Invitrogen). After transfection for 24 hours, relative luciferase activity was calculated by normalizing firefly luminescence to Renilla luminescence using a dual-luciferase reporter assay (Promega) according to the manufacturer's instructions (TECAN-Spark Multifunctional Enzyme Marker).
Statistical analyses
Statistical analyses were performed using two-tailed Student t test and GraphPad Prism 5 software (GraphPad Software). Tumor growth kinetics was assessed using two-way ANOVA. The Mann–Whitney U test was used to determine significance between healthy individuals and patients. A 95% confidence interval and P < 0.05 was considered significant. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
Results
IL6 upregulates expression of Olfr29-ps1 in MDSCs
To identify lncRNA(s) that may regulate the function and differentiation of MDSCs, we found that the lncRNA Olfr29-ps1 was upregulated in MDSCs induced by GM-CSF and IL6 compared with GM-CSF alone (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104718; ref. 20; Fig. 1A). QRT-PCR further confirmed the higher expression of Olfr29-ps1 in GM-CSF– and IL6-induced MDSCs (Fig. 1B). This lncRNA was distributed in the cytoplasm and nuclei of cells (Fig. 1C). Olfr29-ps1 was regulated by proinflammatory cytokine IL6 and tumor-associated factors (Fig. 1D; Supplementary Fig. S1A and S1B). IL6-mediated expression of Olfr29-ps1 was time- and dose-dependent (Supplementary Fig. S1A and S1B). Olfr29-ps1 was significantly decreased in MDSCs from B16 tumor tissues of IL6 knockout mice (Supplementary Fig. S1C). Tumor-associated factor–mediated Olfr29-ps1 expression was suppressed by both STAT3 and JAK3 inhibitors (Supplementary Fig. S1D), indicating that the inflammatory cytokine IL6 can upregulate Olfr29-ps1 expression.
The expression of lncRNA Olfr29-ps1 in MDSCs. A, Heat map of the LncRNA microarray (GSE104718) of MDSCs induced by GM-CSF (40 ng/mL) or GM-CSF (40 ng/mL) plus IL6 (40 ng/mL). The red arrow indicates the lncRNA Olfr29-ps1. B, qRT-PCR of Olfr29-ps1 in MDSCs induced by GM-CSFIL6, or GM-CSF plus IL6. BMC: control. C, Immunostaining and RNA-FISH in MDSCs (C2) before and (C3) after GM-CSF and IL6. C1: control probe. Scale bar, 20 μm. D, qRT-PCR of Olfr29-ps1 in MDSCs induced by GM-CSF, GM-CSF plus IL6, GM-CSF plus B16 tumor supernatant or GM-CSF plus ID8 tumor supernatant for 4 days. E, RT-PCR of Olfr29-ps1 in MDSCs and macrophages, and OR1F2P in U937. Water was used a control (Ctrl). F, Flow cytometry of CD3−HLA-DR−CD33+CD11b+ MDSCs in healthy individuals (n = 8) and patients with colon (n = 12) and/or rectal cancer (n = 8). G, qRT-PCR of OR1F2P in human peripheral blood mononuclear cells from colon cancer (n = 12) or healthy persons (n = 8). H, Immunostaining and RNA-FISH in the human colon cancer (top) and in the peripheral blood in the patients with colon cancer (bottom). Scale bar, 100 μm. Two-tailed Student t test was used in B and D; error bars, SEM; Mann–Whitney U test used in G; three independent experiments in B–E were performed. *, P < 0.05; **, P < 0.01; ***, P < 0.005.
The expression of lncRNA Olfr29-ps1 in MDSCs. A, Heat map of the LncRNA microarray (GSE104718) of MDSCs induced by GM-CSF (40 ng/mL) or GM-CSF (40 ng/mL) plus IL6 (40 ng/mL). The red arrow indicates the lncRNA Olfr29-ps1. B, qRT-PCR of Olfr29-ps1 in MDSCs induced by GM-CSFIL6, or GM-CSF plus IL6. BMC: control. C, Immunostaining and RNA-FISH in MDSCs (C2) before and (C3) after GM-CSF and IL6. C1: control probe. Scale bar, 20 μm. D, qRT-PCR of Olfr29-ps1 in MDSCs induced by GM-CSF, GM-CSF plus IL6, GM-CSF plus B16 tumor supernatant or GM-CSF plus ID8 tumor supernatant for 4 days. E, RT-PCR of Olfr29-ps1 in MDSCs and macrophages, and OR1F2P in U937. Water was used a control (Ctrl). F, Flow cytometry of CD3−HLA-DR−CD33+CD11b+ MDSCs in healthy individuals (n = 8) and patients with colon (n = 12) and/or rectal cancer (n = 8). G, qRT-PCR of OR1F2P in human peripheral blood mononuclear cells from colon cancer (n = 12) or healthy persons (n = 8). H, Immunostaining and RNA-FISH in the human colon cancer (top) and in the peripheral blood in the patients with colon cancer (bottom). Scale bar, 100 μm. Two-tailed Student t test was used in B and D; error bars, SEM; Mann–Whitney U test used in G; three independent experiments in B–E were performed. *, P < 0.05; **, P < 0.01; ***, P < 0.005.
Olfr29-ps1 is located on mouse chromosome 4 and is a pseudogene with a length of 963 bp (Supplementary Fig. S2A and S2B). The sequence of Olfr29-ps1 in vertebrates is conserved (Supplementary Fig. S2A), and the similarity between the Olfr29-ps1 and human OR1F2P sequence is 43%. Human lncRNA OR1F2P was also detected in the human monocyte cell line U937 (Fig. 1E). Its expression was significantly higher in human peripheral blood mononuclear cells (HLADR−CD3−CD11b+CD33+ cells) from colon and rectal cancer patients than in those from healthy individuals (Fig. 1F–H). The mouse and human lncRNA had enrichment of histone modification markers (Supplementary Fig. S3A and S3B) and had no coding capacity (Supplementary Fig. S3C–S3E). These data suggest that Olfr29-ps1 plays a role in the differentiation and function of MDSCs.
Olfr29-ps1 promotes differentiation of monocytic MDSCs
To investigate the effects of Olfr29-ps1 on the differentiation and function of MDSCs, BMCs were first transduced with Olfr29-ps1 shRNA or Olfr29-ps1 lentivirus with a high transduction rate (85%–90%; Supplementary Fig. S4A–S4D) and then cultured in vitro for 4 days in the presence of GM-CSF and IL6. Olfr29-ps1 knockdown decreased the percentage of CD11b+Gr1+ cells significantly at 96 hours (Fig. 2A). Further studies showed that Olfr29-ps1 knockdown reduced the percentage of monocytic (Mo)-MDSCs and increased PMN-MDSCs (Fig. 2B). These effects were also observed in Olfr29-ps1 siRNA-transfected MDSCs (Fig. 2C–E). The overexpression of Olfr29-ps1 increased the percentage of CD11b+Gr1+ cells and Mo-MDSCs, but it impeded the differentiation of PMN-MDSCs (Fig. 2F and G). These results indicated that Olfr29-ps1 is involved in the differentiation of MDSCs and their subsets. To further confirm the effects of Olfr29-ps1 on the differentiation of Mo-MDSCs, we next used a mouse CD45.1+ BMC chimera model. CD45.1 mouse BMCs from homogeneous mice were transduced with Olfr29-ps1 shRNA or Olfr29-ps1/lentivirus and then injected into WT mice via the tail vein following the indicated timeline (Fig. 2H). CD45.1+ cells were detected in the spleen at day 1 after injecting lentivirus-transduced BMCs (Fig. 2I), indicating successful establishment of the chimera model. In mice injected by exogenous Olfr29-ps1–transduced BMCs, the proportion of Mo-MDSCs significantly increased in the spleen of mice, whereas decreased Mo-MDSCs were seen in the spleen of mice injected with Olfr29-ps1–knockdown BMCs (Fig. 2J). Taken together, our data demonstrated that Olfr29-ps1 promotes the differentiation of Mo-MDSCs.
Olfr29-ps1 promotes the differentiation of Mo-MDSCs. A, Flow-cytometric analysis of MDSCs transduced with control shRNA/lentiviruses (Lv-shNC) or Olfr29-ps1 shRNA/lentiviruses (Lv-shOlfr29-ps1) at 48 and 96 hours. B, Flow-cytometric analysis of CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSCs transduced with control shRNA/lentiviruses (Lv-shNC) or Olfr29-ps1 shRNA/Lentiviruses (Lv-shOlfr29-ps1) at 48 and 96 hours. C, qRT-PCR of Olfr29-ps1 in MDSCs transfected with siNC (siRNA control) or siOlfr29-ps1 (Olfr29-ps1 siRNA). D and E, Flow-cytometric analysis of (D) MDSCs or (E) CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSCs transfected with siNC or siOlfr29-ps1. F, Flow-cytometric analysis of MDSCs transduced with empty lentivirus control (Lv-oeNC) or Olfr29-ps1 lentivirus (Lv-oeOlfr29-ps1). G, Flow-cytometric analysis of CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSC subsets after transfection with Lv-oeNC or Lv-oeOlfr29-ps1. H, Schematic of in vivo experiments. After injecting genetically modified CD45.1+ BMCs (1 × 107/mouse), the spleens were checked by confocal microscopy (day 1) and flow cytometry (day 7). I, Representative images of CD45.1+ cells in the spleen of mice indicated in H. NC, isotypic antibody. Scale bar, 50 μm. J, Flow-cytometric analysis of Gr1+CD11b+, CD11b+Ly6G+Ly6C+, and CD11b+Ly6G−Ly6C+ MDSCs in the spleens of mice after infusing Lv-shOlfr29-ps1, Lv-oeOlfr29-ps1, and oeNC and Lv-shNC control lentivirus–treated CD45.1+ MDSCs (6 mice/group). Two-tailed Student t test was used; error bars, SEM; three independent experiments in all panels were performed. NS, no significance; *, P < 0.05; **, P < 0.01.
Olfr29-ps1 promotes the differentiation of Mo-MDSCs. A, Flow-cytometric analysis of MDSCs transduced with control shRNA/lentiviruses (Lv-shNC) or Olfr29-ps1 shRNA/lentiviruses (Lv-shOlfr29-ps1) at 48 and 96 hours. B, Flow-cytometric analysis of CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSCs transduced with control shRNA/lentiviruses (Lv-shNC) or Olfr29-ps1 shRNA/Lentiviruses (Lv-shOlfr29-ps1) at 48 and 96 hours. C, qRT-PCR of Olfr29-ps1 in MDSCs transfected with siNC (siRNA control) or siOlfr29-ps1 (Olfr29-ps1 siRNA). D and E, Flow-cytometric analysis of (D) MDSCs or (E) CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSCs transfected with siNC or siOlfr29-ps1. F, Flow-cytometric analysis of MDSCs transduced with empty lentivirus control (Lv-oeNC) or Olfr29-ps1 lentivirus (Lv-oeOlfr29-ps1). G, Flow-cytometric analysis of CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSC subsets after transfection with Lv-oeNC or Lv-oeOlfr29-ps1. H, Schematic of in vivo experiments. After injecting genetically modified CD45.1+ BMCs (1 × 107/mouse), the spleens were checked by confocal microscopy (day 1) and flow cytometry (day 7). I, Representative images of CD45.1+ cells in the spleen of mice indicated in H. NC, isotypic antibody. Scale bar, 50 μm. J, Flow-cytometric analysis of Gr1+CD11b+, CD11b+Ly6G+Ly6C+, and CD11b+Ly6G−Ly6C+ MDSCs in the spleens of mice after infusing Lv-shOlfr29-ps1, Lv-oeOlfr29-ps1, and oeNC and Lv-shNC control lentivirus–treated CD45.1+ MDSCs (6 mice/group). Two-tailed Student t test was used; error bars, SEM; three independent experiments in all panels were performed. NS, no significance; *, P < 0.05; **, P < 0.01.
Olfr29-ps1 promotes the immunosuppressive function of MDSCs
To analyze the effects of Olfr29-ps1 on the immunosuppressive function of MDSCs, we cocultured ovalbumin (OVA)-specific OT-I or OT-II splenic cells with MDSCs. Although Olfr29-ps1–knockdown MDSCs were added into OT-I CD8+ or OT-II CD4+ T cells, which respond to MHCI- or MHCII-restricted OVA peptides, significantly weakened the immunosuppressive function of MDSCs, whereas Olfr29-ps1/lentivirus–transduced MDSCs exhibited more suppression on IFNγ production compared with control MDSCs (Fig. 3A and B). The inhibition of MDSCs on T cells is dependent on Arg-1, iNOS, NOX2, Cox2, and their products (6). Olfr29-ps1–knockdown cells had lower NO, H2O2, and ROS (Fig. 3C and D), whereas increased NO, H2O2, and ROS were observed in Olfr29-ps–overexpressing MDSCs compared with controls (Fig. 3C and D). Similar effects were also observed in OR1F2P-knockdown or -overexpressing human MDSCs (Supplementary Fig. S5A–S5C). Western blot analysis showed that Olfr29-ps1–silencing reduced protein levels of Arg-1, iNOS, Cox2, and Nox2, whereas the protein levels of Arg-1, iNOS, Cox2, and Nox2 were upregulated in Olfr29-ps1–overexpressing MDSCs (Fig. 3E). Mo-MDSCs can produce high amounts of Arg-1 (28), whereas PMN-MDSCs mainly depend on H2O2 (28). Olfr29-ps1–knockdown Mo-MDSCs had decreased Arg-1, and lower H2O2 was also detected in Olfr29-ps1–knockdown PMN-MDSCs. Exogenous Olfr29-ps1 promoted the production of Arg-1 in Mo-MDSC and H2O2 in PMN-MDSCs (Fig. 3F and G). Both Olfr29-ps1–knockdown Mo-MDSCs and PMN-MDSCs also had a decreased immunosuppressive function, whereas Olfr29-ps1 promoted the immunosuppressive function of both Mo-MDSCs and PMN-MDSCs (Fig. 3H and I). These results indicated that Olfr29-ps1 promotes the immunosuppressive effect of MDSCs.
Olfr29-ps1 promotes the suppressive function of MDSCs in vitro. A and B, The activity of T cells was measured by their capacity to produce IFNγ upon OVA-MHCI– or OVA-MHCII–specific peptide stimulation (10 μg/mL) for 24 hours. IFNγ in the supernatants was detected by ELISA. A, Suppressive capacity of Olfr29-ps1 shRNA/lentivirus (Lv-shOlfr29-ps1)–transduced MDSCs at the indicated ratios. B, Suppressive capacity of Olfr29-ps1/lentivirus (Lv-oeOlfr29-ps1)–transduced MDSCs at the indicated ratios. C, NO and H2O2 production in Lv-shOlfr29-ps1–transduced MDSCs (left) and Lv-oeOlfr29-ps1–transduced MDSCs (right). D, Representative flow-cytometric histograms of ROS in Lv-shOlfr29–transduced MDSCs (left) and Lv-oeOlfr29-ps1–transduced MDSCs (right). E, Immunoblotting of Arg-1, COX2, NOX2, and iNOS in Olfr29-ps1–knockdown MDSCs and exogenous Olfr29-ps1–treated MDSCs. F, Arg-1 activity (left) and H2O2 production (right) in Lv-Olfr29-ps1–transduced Mo-MDSCs. G, Arg-1 activity (left) and H2O2 production (right) in Lv-oeOlfr29-ps1–transduced PMN-MDSCs. H and I, M-MDSCs and PMN-MDSCs were sorted from Lv-shOlfr29-ps1–transduced MDSCs or Lv-oeOlfr29-ps1–transduced MDSCs. The activity of T cells was measured as indicated in A–B. Suppressive capacity of (H) Mo-MDSCs and (I) PMN-MDSCs (T:MDSC ratio 10:1, 10 μg/mL peptides). Supernatants were analyzed after 24 hours. Two-tailed Student t test was used. Error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Three independent experiments in all panels were performed.
Olfr29-ps1 promotes the suppressive function of MDSCs in vitro. A and B, The activity of T cells was measured by their capacity to produce IFNγ upon OVA-MHCI– or OVA-MHCII–specific peptide stimulation (10 μg/mL) for 24 hours. IFNγ in the supernatants was detected by ELISA. A, Suppressive capacity of Olfr29-ps1 shRNA/lentivirus (Lv-shOlfr29-ps1)–transduced MDSCs at the indicated ratios. B, Suppressive capacity of Olfr29-ps1/lentivirus (Lv-oeOlfr29-ps1)–transduced MDSCs at the indicated ratios. C, NO and H2O2 production in Lv-shOlfr29-ps1–transduced MDSCs (left) and Lv-oeOlfr29-ps1–transduced MDSCs (right). D, Representative flow-cytometric histograms of ROS in Lv-shOlfr29–transduced MDSCs (left) and Lv-oeOlfr29-ps1–transduced MDSCs (right). E, Immunoblotting of Arg-1, COX2, NOX2, and iNOS in Olfr29-ps1–knockdown MDSCs and exogenous Olfr29-ps1–treated MDSCs. F, Arg-1 activity (left) and H2O2 production (right) in Lv-Olfr29-ps1–transduced Mo-MDSCs. G, Arg-1 activity (left) and H2O2 production (right) in Lv-oeOlfr29-ps1–transduced PMN-MDSCs. H and I, M-MDSCs and PMN-MDSCs were sorted from Lv-shOlfr29-ps1–transduced MDSCs or Lv-oeOlfr29-ps1–transduced MDSCs. The activity of T cells was measured as indicated in A–B. Suppressive capacity of (H) Mo-MDSCs and (I) PMN-MDSCs (T:MDSC ratio 10:1, 10 μg/mL peptides). Supernatants were analyzed after 24 hours. Two-tailed Student t test was used. Error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Three independent experiments in all panels were performed.
To further investigate the effects of Olfr29-ps1 on the differentiation and function of MDSCs in vivo, we used a murine B16 melanoma model. Olfr29-ps1–knockdown or overexpressing CD45.1+ MDSCs were injected into the mice after inoculating B16 tumor cells, and then the tumor growth was monitored. Compared with the control group, Olfr29-ps1 knockdown decreased the immunosuppressive function of CD45.1+ MDSCs, and tumors grew slower in mice injected with Olfr29-ps1–knockdown MDSCs (Fig. 4A). A smaller tumor volume and lighter tumor weight were detected in these mice (Fig. 4B and C), whereas Olfr29-ps1–overexpressing MDSCs caused faster tumor growth, larger tumor volume, and heavier tumor weight compared with control mice (Fig. 4A–C). The mice injected with Olfr29-ps1 shRNA/lentivirus–transduced MDSCs had more CD4+ and CD8+ T cells in the tumor tissues compared with the control group, whereas fewer CD4+ and CD8+ T cells appeared in the tumor tissue of the mice injected with Olfr29-ps1/lentivirus–transduced MDSCs (Fig. 4D). These results indicated that Olfr29-ps1 enhances the inhibition ability of MDSCs. The proportion of the CD11b+Ly6G−Ly6Chi subset was reduced in the tumor of mice injected with Olfr29-ps1–knockdown CD45.1+ MDSCs, whereas this subset increased in the tumor with Olfr29-ps1–overexpressing CD45.1+ MDSCs (Fig. 4E), further confirming that Olfr29-ps1 promoted differentiation of Mo-MDSCs. CD45.1+ cells were also detected in the tumor tissues and spleen in these tumor-bearing mice (Fig. 4F). Taken together, these results support our findings that Olfr29-ps1 not only promotes the immunosuppressive function but also the differentiation of Mo-MDSCs.
Olfr29-ps1 promotes differentiation and suppressive function of MDSCs in vivo. Tumor growth (A), tumor size (B), and tumor weight (C) in C57/BL6 mice bearing B16 tumors (N = 6/group) injected with CD45.1+ MDSCs transduced with Olfr29-ps1 shRNA/lentivirus (Lv-shOlfr29-ps1) or Olfr29-ps1/lentivirus (Lv-oeOlfr29-ps1). oeNC and Lv-shNC: overexpressing and knockdown control lentiviruses, respectively. D, Flow-cytometric analysis of CD4+ or CD8+ T cells in the tumors of mice injected with CD45.1+MDSCs transduced with Lv-shOlfr29-ps1 or Lv-oeOlfr29-ps1. E, Flow-cytometric analysis of CD45.1+ Gr1+CD11b+, Gr1highCD11b+, and Gr1lowCD11b+ MDSCs in tumors of mice bearing B16 tumors after infusing Lv-shOlfr29-ps1, Lv-oeOlfr29-ps1, and oeNC and Lv-shNC control lentiviruses treated CD45.1+ MDSCs. F, The representative images of CD45.1+ cells by confocal microscopy in the tumor site and spleen. NC, isotype antibody. Scale bar, 50 μm. Two-way ANOVA was used in A; the Mann–Whitney U test was used in C; Two-tailed Student t test was used in D–F; error bars, SD; *, P < 0.05; **, P < 0.01; ***, P < 0.005. One representative of three experiments.
Olfr29-ps1 promotes differentiation and suppressive function of MDSCs in vivo. Tumor growth (A), tumor size (B), and tumor weight (C) in C57/BL6 mice bearing B16 tumors (N = 6/group) injected with CD45.1+ MDSCs transduced with Olfr29-ps1 shRNA/lentivirus (Lv-shOlfr29-ps1) or Olfr29-ps1/lentivirus (Lv-oeOlfr29-ps1). oeNC and Lv-shNC: overexpressing and knockdown control lentiviruses, respectively. D, Flow-cytometric analysis of CD4+ or CD8+ T cells in the tumors of mice injected with CD45.1+MDSCs transduced with Lv-shOlfr29-ps1 or Lv-oeOlfr29-ps1. E, Flow-cytometric analysis of CD45.1+ Gr1+CD11b+, Gr1highCD11b+, and Gr1lowCD11b+ MDSCs in tumors of mice bearing B16 tumors after infusing Lv-shOlfr29-ps1, Lv-oeOlfr29-ps1, and oeNC and Lv-shNC control lentiviruses treated CD45.1+ MDSCs. F, The representative images of CD45.1+ cells by confocal microscopy in the tumor site and spleen. NC, isotype antibody. Scale bar, 50 μm. Two-way ANOVA was used in A; the Mann–Whitney U test was used in C; Two-tailed Student t test was used in D–F; error bars, SD; *, P < 0.05; **, P < 0.01; ***, P < 0.005. One representative of three experiments.
Olfr29-ps1–mediated effects on MDSCs depend on m6A modification
N6-methyladenosine (m6A) is the most abundant internal modification in eukaryotic messenger RNAs (mRNA) and lncRNAs. This modification is cell type– and condition-dependent and is reversible (29, 30). The formation of m6A needs methyltransferase, including METTL3 (31), METTL14 (32), and WTAP (Wilms tumor 1–associated protein; ref. 33). Studies have shown that lncRNAs can be modified by m6A (34, 35), and the most prevalent m6A consensus sequence is GGACT (36). LncRNA Olfr29-ps1 has seven conserved sequences of GGAC that can potentially be modified by m6A (Fig. 5A). RIP-PCR showed that Olfr29-ps1 was modified by m6A in MDSCs induced by GM-CSF plus IL6 (Fig. 5B). Because m6A modification may affect the function of mRNAs, such as mRNA splicing, transport, stabilization, and immune tolerance (30, 32, 37, 38), we hypothesized that m6A modification could promote the formation and stability of Olfr29-ps1. We, thus, investigated the effects of m6A modification on Olfr29-ps1 expression. METTL3 is a methyltransferase required for the formation of m6A (32). Silencing METTL3 reduced Olfr29-ps1 expression in MDSCs, whereas overexpression of METTL3 increased Olfr29-ps1 in MDSCs (Fig. 5C), suggesting that m6A modification promoted the formation and stability of Olfr29-ps1. Further studies showed that the silencing METTL3 reduced the percentage of the CD11b+Ly6G−Ly6Chi subset (Fig. 5D), but overexpression of METTL3 had the opposite effect (Fig. 5E), suggesting that Olfr29-ps1–mediated Mo-MDSC differentiation is dependent on m6A modification. METTL3 knockdown in MDSCs also attenuated the immunosuppressive function of MDSCs on OT-I and OT-II T cells, whereas the overexpression of METTL3 enhanced the immunosuppressive effect of MDSCs on these T cells (Fig. 5F), indicating that Olfr29-ps1–mediated immunosuppression also depended on the m6A. Thus, these results demonstrated that Olfr29-ps1–mediated immunosuppressive function and Mo-MDSC differentiation depends on the m6A modification.
The effect of METTL3-mediated m6A modification on the suppressive function and differentiation of MDSCs. A, Motifs that can be modified by m6A in Olfr29-ps1. Sequences indicated in blue. B, RIP-PCR of Olfr29-ps1 using anti-m6A in MDSCs induced by GM-CSF, IL6, or GM-CSF plus IL6. PC, positive control (Olfr29-ps1 plasmids); IgG, isotype control. C, qRT-PCR of Olfr29-ps1 in MDSCs transfected with METTL3 siRNAs (siMETTL3; left) or pcDNA-3.1-METTL3 plasmids (oeMETTL3; right). D, Representative flow cytometric (left) and group (right) analyses of CD11b+Gr1+ MDSCs transfected with siMETTL3 or oeMETTL3. E, Representative flow cytometric (left) and group (right) analyses of CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSC subsets transfected with siMETTL3 or oeMETTL3. F, Suppressive capacity of siMETTL3- or oeMETTL3-transfected MDSCs. The activity of T cells was measured by their capacity to produce IFNγ upon OVA-MHCI– or OVA-MHCII–specific peptide stimulation (T:MSDC ratio 10:1, 10 μg peptides/mL). IFNγ in the supernatants was detected after 24 hours by ELISA. SiNC (control siRNA) and oeNC (exogenous control plasmids) were used as controls. Two-tailed Student t test was used; error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Three independent experiments in all panels were performed.
The effect of METTL3-mediated m6A modification on the suppressive function and differentiation of MDSCs. A, Motifs that can be modified by m6A in Olfr29-ps1. Sequences indicated in blue. B, RIP-PCR of Olfr29-ps1 using anti-m6A in MDSCs induced by GM-CSF, IL6, or GM-CSF plus IL6. PC, positive control (Olfr29-ps1 plasmids); IgG, isotype control. C, qRT-PCR of Olfr29-ps1 in MDSCs transfected with METTL3 siRNAs (siMETTL3; left) or pcDNA-3.1-METTL3 plasmids (oeMETTL3; right). D, Representative flow cytometric (left) and group (right) analyses of CD11b+Gr1+ MDSCs transfected with siMETTL3 or oeMETTL3. E, Representative flow cytometric (left) and group (right) analyses of CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSC subsets transfected with siMETTL3 or oeMETTL3. F, Suppressive capacity of siMETTL3- or oeMETTL3-transfected MDSCs. The activity of T cells was measured by their capacity to produce IFNγ upon OVA-MHCI– or OVA-MHCII–specific peptide stimulation (T:MSDC ratio 10:1, 10 μg peptides/mL). IFNγ in the supernatants was detected after 24 hours by ELISA. SiNC (control siRNA) and oeNC (exogenous control plasmids) were used as controls. Two-tailed Student t test was used; error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Three independent experiments in all panels were performed.
Olfr29-ps1–mediated effects on MDSCs is through sponging miR-214-3p
We next investigated how Olfr29-ps1 induces immunosuppressive function and Mo-MDSC differentiation. Studies have shown that lncRNAs can competitively bind to microRNAs in the cytoplasm by acting as a competing endogenous RNA (ceRNA) and, thereby, regulate cell differentiation and function (23). Because Olfr29-ps1 was located in the cytoplasm, we speculated that Olfr29-ps1–mediated immunosuppressive function and Mo-MDSC differentiation could be through ceRNAs. RNAHybrid program and miRDB showed that Olfr29-ps1 potentially interacted with multiple miRNAs, such as miR-214-3p. MiR-214-3p was significantly increased in Olfr29-psl–knockdown MDSCs (Fig. 6A). Olfr29-ps1 also had two potential binding sites for miR-214-3p (Fig. 6B). To determine whether Olfr29-ps1–mediated effects on MDSCs were through sponging miR-214-3p, we performed a dose–response experiment. The expression of miR-214-3p was gradually upregulated with increasing Olfr29-ps1 shRNA/lentivirus concentration, and while Olfr29-ps1 increased, decreased miR-214-3p was also observed (Supplementary Fig. S6A–S6C). When miR-214-3p gradually decreased with time during MDSC differentiation in vitro, the expression of Olfr29-ps1 was upregulated (Supplementary Fig. S6D). Transfection of miR-214-3p did not significantly affect the expression of Olfr29-ps1 (Supplementary Fig. S6E). Taken together, these findings demonstrated that there exists a negative correlation between Olfr29-ps1 and miR-214-3p.
Olfr29-ps1 regulates MDSC differentiation by miR-214-3p. A, qRT-PCR of miR-214-3p, miR-149-5p, and miR-361-3p in MDSCs transduced with Olfr29-ps1 shRNA/lentivirus (Lv-shOlfr29-ps1). B, Thermodynamic energy prediction for the association of Olfr29-ps1 and miR-214-3p by RNAHybrid program. Partial sequences of Olfr29-ps1 (top) and miR-214-3p (bottom) are shown. Numbers above the sequences indicate the positions of nucleotides relative to the transcriptional start site of Olfr29-ps1. C, Dual-luciferase reporter assay of 293T cells cotransfected luc-Olfr29-ps1, miR-214-3p mimic (miR-214-3p), or mimic control (miR-NC; left) or MDSCs cotransfected luc-Olfr29-ps1 and miR-214-3p inhibitor or inhibitor control (miR-NC; right). D, Luciferase activity of luc-Olfr29-ps1 after the addition of miR-214-3p mimic in MDSCs transfected with METTL3 siRNAs (siMETTL3; left) and pcDNA-3.1-METTL3 (oeMETTL3; right). E, Coprecipitation of Olfr29-ps1 and miRNAs associated with Ago2. Anti-Ago2 RIP was performed in MDSC lysates. IgG, control. F and G, Biotin-labeled RNA pulldown experiments in MDSCs. MDSC lysates were incubated with in vitro–synthesized biotin-labeled Olfr29-ps1 sense or antisense RNA, followed by qRT-PCR to detect (F) miRNAs and (G) Western blotting to detect Ago2 associated with Olfr29-ps1. Bio-Olfr29-ps1-AS: biotinylated Olfr29-ps1-antisense RNA; bio-Olfr29-ps1-S: biotinylated Olfr29-ps1-sense RNA. H, Representative flow cytometric and (I) group analyses of CD11b+Gr1+, CD11b+Ly6G+Ly6C+, and CD11b+Ly6G−Ly6C+ MDSCs. J, Suppressive capacity of MDSCs. The activity of T cells was measured by their capacity to produce IFNγ upon OVA-MHCI– or OVA-MHCII–specific peptide stimulation (T:MDSC ratio 10:1, 10 μg peptides/mL). IFNγ in the supernatants was detected by ELISA after 24 hours. The MDSCs were induced by transfecting BMCs with shNC, shOlfr29-ps1 and mimic NC, shOlfr29-ps1 and miR-214-3p mimic, shOlfr29-ps1 and inhibitor NC, or shOlfr29-ps1 and miR-214-3p inhibitor, respectively. mimic NC: mimics control; inhibitor NC: inhibitor control. Two-tailed Student t test was used; error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005; NS, no significance. Three independent experiments in all panels were performed.
Olfr29-ps1 regulates MDSC differentiation by miR-214-3p. A, qRT-PCR of miR-214-3p, miR-149-5p, and miR-361-3p in MDSCs transduced with Olfr29-ps1 shRNA/lentivirus (Lv-shOlfr29-ps1). B, Thermodynamic energy prediction for the association of Olfr29-ps1 and miR-214-3p by RNAHybrid program. Partial sequences of Olfr29-ps1 (top) and miR-214-3p (bottom) are shown. Numbers above the sequences indicate the positions of nucleotides relative to the transcriptional start site of Olfr29-ps1. C, Dual-luciferase reporter assay of 293T cells cotransfected luc-Olfr29-ps1, miR-214-3p mimic (miR-214-3p), or mimic control (miR-NC; left) or MDSCs cotransfected luc-Olfr29-ps1 and miR-214-3p inhibitor or inhibitor control (miR-NC; right). D, Luciferase activity of luc-Olfr29-ps1 after the addition of miR-214-3p mimic in MDSCs transfected with METTL3 siRNAs (siMETTL3; left) and pcDNA-3.1-METTL3 (oeMETTL3; right). E, Coprecipitation of Olfr29-ps1 and miRNAs associated with Ago2. Anti-Ago2 RIP was performed in MDSC lysates. IgG, control. F and G, Biotin-labeled RNA pulldown experiments in MDSCs. MDSC lysates were incubated with in vitro–synthesized biotin-labeled Olfr29-ps1 sense or antisense RNA, followed by qRT-PCR to detect (F) miRNAs and (G) Western blotting to detect Ago2 associated with Olfr29-ps1. Bio-Olfr29-ps1-AS: biotinylated Olfr29-ps1-antisense RNA; bio-Olfr29-ps1-S: biotinylated Olfr29-ps1-sense RNA. H, Representative flow cytometric and (I) group analyses of CD11b+Gr1+, CD11b+Ly6G+Ly6C+, and CD11b+Ly6G−Ly6C+ MDSCs. J, Suppressive capacity of MDSCs. The activity of T cells was measured by their capacity to produce IFNγ upon OVA-MHCI– or OVA-MHCII–specific peptide stimulation (T:MDSC ratio 10:1, 10 μg peptides/mL). IFNγ in the supernatants was detected by ELISA after 24 hours. The MDSCs were induced by transfecting BMCs with shNC, shOlfr29-ps1 and mimic NC, shOlfr29-ps1 and miR-214-3p mimic, shOlfr29-ps1 and inhibitor NC, or shOlfr29-ps1 and miR-214-3p inhibitor, respectively. mimic NC: mimics control; inhibitor NC: inhibitor control. Two-tailed Student t test was used; error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005; NS, no significance. Three independent experiments in all panels were performed.
Next, we determined whether Olfr29-ps1 could interact with miR-214-3p. We constructed a dual-luciferase reporter plasmid containing the Olfr29-ps1 sequence (luc-Olfr29-ps1), and then cotransfected luc-Olfr29-ps1 and miR-214-3p mimics into the 293T cell. MiR-214-3p mimics significantly reduced the luciferase activity of luc-Olfr29-ps1, whereas miR-214-3p inhibition enhanced relative luciferase activity of luc-Olfr29-ps1 (Fig. 6C), indicating that Olfr29-ps1 could interact with miR-214-3p. Next, we also determined whether m6A modification affected the interaction of Olfr29-ps1 and miR-214-3p. Silencing METTL3 significantly reduced the inhibitory function of miR-214-3p on the luciferase activity of luc-Olfr29-ps1 (Fig. 6D). Conversely, the inhibition of miR-214-3p on the luciferase activity of luc-Olfr29-ps1 was promoted by overexpressed METTL3 (Fig. 6D), indicating that the m6A modification of Olfr29-ps1 is necessary for the interaction between Olfr29-ps1 and miR-214-3p. To further validate the direct binding between miR-214-3p and Olfr29-ps1 at endogenous levels, we performed an anti-Ago2 RNA immunoprecipitation in MDSC extracts. As expected, both miR-214-3p and Olfr29-ps1 were specifically enriched in the Ago2 complexes (Fig. 6E). A RNA–protein pulldown assay further validated the specific association between miR-214-3p and Olfr29-ps1 (Fig. 6F). Ago2 enrichment was also observed in the Olfr29-ps1 pulldown complex (Fig. 6G), indicating that Olfr29-ps1 was recruited to Ago2-related RNA complexes and functionally interacts with miR-214-3p. We finally tested whether Olfr29-ps1–mediated suppressive function and differentiation of MDSCs was through miR-214-3p. MiR-214-3p mimics weakened Olfr29-ps1–mediated immunosuppressive function and impeded the differentiation of Mo-MDSCs, whereas these functions were strengthened by miR-214-3p inhibition (Fig. 6H–J). Taken together, our data indicated that Olfr29-ps1 may regulate immunosuppressive function and Mo-MDSC differentiation through sponging miR-214-3p.
Olfr29-ps1–mediated effects on MDSCs are through increased MyD88
MicroRNAs regulate cell processes through regulating target gene expression. MiR-214-3p can regulate the expression of MyD88 (39). QRT-PCR and Western blotting showed that miR-214-3p reduced mRNA and protein levels of MyD88 (Fig. 7A), whereas miR-214-3p inhibition increased its expression (Fig. 7B), suggesting that the modulation of miR-214-3p on MDSC differentiation may be through downregulating MyD88. The expression of MyD88 in MDSCs was downregulated by the Olfr29-ps1 shRNA (Fig. 7C), but Olfr29-ps1 increased mRNA and protein levels of MyD88 (Fig. 7D). The expression patterns of Olfr29-ps1, miR-214-3p, and MyD88 in tumor MDSCs from different mice were the same as those in the MDSCs induced in vitro (Fig. 7E–G). These results indicate that Olfr29-ps1 is positively correlated with MyD88 expression. Previous studies have shown that the deletion of MyD88 can regulate the differentiation of MDSCs (40, 41). MyD88 knockdown affected the immunosuppressive function of MDSCs and the differentiation of Mo-MDSCs (Fig. 7H; Supplementary Fig. S7A–S7D). Thus, our data indicated that Olfr29-ps1 mediated the differentiation and functions of MDSCs are through the release of MyD88 after sponging miR-214-3p.
Olfr29-ps1 regulates Myd88 expression by sponging miR-214-3p. A, qRT-PCR and immunoblotting analyses of MyD88 in MDSCs transfected by mimic NC or miR-214-3p mimic. B, qRT-PCR and immunoblotting analyses of MyD88 in MDSCs transfected by inhibitor NC or miR-214-3p inhibitor. C, qRT-PCR and immunoblotting of MyD88 in MDSCs transduced with Olfr29-ps1 shRNA/lentiviruses (Lv-shOlfr29-ps1). D, qRT-PCR and immunoblotting of MyD88 in MDSCs transfected with Olfr29-ps1/lentiviruses (Lv-oeOlfr29-ps1). E–G, qRT-PCR of (E) Olfr29-ps1, (F) miR-214-3p, and (G) MyD88 in tumor MDSCs from mice injected with CD45.1+ MDSCs transduced with Lv-shOlfr29-ps1 or Olfr29-ps1/lentiviruses (Lv-oeOlfr29-ps1; 6 mice/group). H, Representative flow cytometric (left) and group (right) analyses of MDSCs transfected with MyD88 siRNA (siMyD88). SiNC: siRNA control. Two-tailed Student t test was used in A–H; error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Three independent experiments in all panels were performed.
Olfr29-ps1 regulates Myd88 expression by sponging miR-214-3p. A, qRT-PCR and immunoblotting analyses of MyD88 in MDSCs transfected by mimic NC or miR-214-3p mimic. B, qRT-PCR and immunoblotting analyses of MyD88 in MDSCs transfected by inhibitor NC or miR-214-3p inhibitor. C, qRT-PCR and immunoblotting of MyD88 in MDSCs transduced with Olfr29-ps1 shRNA/lentiviruses (Lv-shOlfr29-ps1). D, qRT-PCR and immunoblotting of MyD88 in MDSCs transfected with Olfr29-ps1/lentiviruses (Lv-oeOlfr29-ps1). E–G, qRT-PCR of (E) Olfr29-ps1, (F) miR-214-3p, and (G) MyD88 in tumor MDSCs from mice injected with CD45.1+ MDSCs transduced with Lv-shOlfr29-ps1 or Olfr29-ps1/lentiviruses (Lv-oeOlfr29-ps1; 6 mice/group). H, Representative flow cytometric (left) and group (right) analyses of MDSCs transfected with MyD88 siRNA (siMyD88). SiNC: siRNA control. Two-tailed Student t test was used in A–H; error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005. Three independent experiments in all panels were performed.
We also found that two other miRNAs, not miR-214-3p, were capable of binding to Olfr29-ps1 (Supplementary Fig. S8A–S8D). This may explain why the effects seen by manipulating the Olfr29-ps1 pseudogene and miRNAs were not significant. Another phenomenon was that although the Olfr29-ps1 pseudogene had a minor effect on Mo-MDSCs, it could affect tumor growth. This dichotomy was derived from tumor-mediated miR-214-3p (Supplementary Fig. S9A–S9E).
Discussion
In this study, we identified the pseudogene lncRNA Olfr29-ps1, which may promote the immunosuppressive function and differentiation of Mo-MDSCs. Olfr29-ps1 can sponge miR-214-3p to cause increased expression of MyD88, a target gene of miR-214-3p. The interaction of Olfr29-ps1 and miR-214-3p is dependent on the m6A modification of Olfr29-ps1. We also found that Olfr29-ps1 was expressed in tumor MDSCs, suggesting a potential role of Olfr29-ps1 in antitumor immunity. Thus, our data demonstrated an m6A-modified Olfr29-ps1/miR-214-3p/MyD88 network to regulate the immunosuppressive function and differentiation of Mo-MDSCs in the inflammatory tumor environment.
Multiple data have described the presence of MDSCs in patients with tumors, such as colon cancer, lung cancer, breast cancer, pancreatic adenocarcinomas, urothelial carcinoma, kidney cancer, and glioblastoma (28, 42). Studies have shown that Mo-MDSCs have higher suppressive activity than PMN-MDSCs (43). In tumor tissues, Mo-MDSCs are more prominent than PMN-MDSCs, and Mo-MDSCs may rapidly differentiate into tumor-associated macrophages (28). Mo-MDSCs can produce high amounts of NO, Arg-1, and immune-suppressive cytokines, which have longer half-lives than the ROS produced by PMN-MDSCs (43). Less immunosuppressive PMN-MDSCs than Mo-MDSCs have also been confirmed at the single-cell level (28, 42). PMN-MDSCs also have a short half-life. Thus, our data provide insights that could help to develop novel treatments for tumors through modulating the expression of Olfr29-ps1 to control the differentiation of MDSCs into Mo-MDSCs.
Studies found that there exists a novel class of lncRNAs transcribed from pseudogenes with more than 200 nucleotides, which are called the pseudogene lncRNAs (14, 23). Some pseudogene lncRNAs are demonstrated to control ancestral gene expression by acing as ceRNAs to sponge miRNAs, altering the stability of the ancestral mRNA or affecting the promoter activity of ancestral genes (44, 45). Ancestral genes have defects within the evolution of the genome, such as lack of promoters, premature termination codons, or code-shifting mutations, resulting in pseudogenes. However, some pseudogene lncRNAs can affect other gene expression. For example, Lethe, induced by TNFα, negatively regulates NF-κB activity by binding to NF-κB–RelA to fine tune the inflammatory response (22). In this study, we found the pseudogene lncRNA Olfr29-ps1 could regulate the immunosuppressive function and differentiation of Mo-MDSCs through competitively binding to miR-214-3p, thereby releasing the expression of its target gene MyD88 in response to inflammatory factors.
The in vitro effects of the pseudogene and miRNA expression were not robust and similar to those seen in vivo. LncRNAs generally exert their function through multiple mechanisms, such as interaction with miRNAs, which can be differentially regulated in different environments. Thus, the environmental differences between the in vitro tissue culture and the in vivo tumors could affect the function of lncRNAs, including pseudogene Olfr29-ps1. Future experiments will need to be conducted to more fully understand how lncRNAs function in different environments and different experimental setups.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: R. Yang, W. Shang
Development of methodology: W. Shang
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.):W. Shang, Y. Gao
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): W. Shang
Writing, review, and/or revision of the manuscript: R. Yang, W. Shang
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Z. Tang, Y. Zhang
Study supervision: R. Yang, Y. Zhang
Acknowledgments
This research was supported by the National Key Research and Development Program of China (2016YFC1303604) and NSFC grants 91842302, 91029736, 9162910, 81600436, and 91442111, the Joint NSFC-ISF Research Program, and the State Key Laboratory of Medicinal Chemical Biology.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.