Myeloid-derived suppressor cells (MDSCs) are pathologically activated neutrophils and monocytes with potent immunosuppressive activity that regulate immune responses in the tumor microenvironment. We identified a novel long noncoding RNA (lncRNA), named as lnc57Rik, in the MDSCs that controls their immunosuppressive functions. Lnc57Rik was induced in in vitro and in vivo inflammatory settings and upregulated the genes related to MDSC-mediated immunosuppression, including Arg-1, NOS2, NOX2, and COX2. Furthermore, Lnc57Rik can not only bind with the C/EBPβ isoform liver-enriched activator protein to activate C/EBPβ but also with the methyltransferase WD repeat-containing protein 5 that enables the enrichment of histone H3 trimethylated lysine 4 marks on the promoter regions of Arg-1, NOS2, NOX2, and COX2, eventually resulting in their transcriptional activation. Furthermore, the conserved human lnc57Rik has a similar function as murine lnc57Rik. Taken together, upregulation of lnc57Rik in the tumor microenvironment promotes the immunosuppressive function of MDSCs.
Myeloid-derived suppressor cells (MDSCs) are pathologically activated neutrophils and monocytes with potent immunosuppressive activity that are closely associated with poor clinical outcomes in cancers (1–4) and significantly limit the effects of cancer therapies (5, 6). The murine MDSCs are broadly immunophenotyped as CD11b+Gr-1+ and classified into the CD11b+Ly6Cint/lowLy6G+ polymorphonuclear MDSCs (PMN-MDSCs) and CD11b+Ly6ChiLy6G− monocytic MDSCs (7, 8). Human monocytic (CD14+) and PMN (CD15+) MDSCs are described as lineage-negative cells that coexpress CD11b and CD33 but lack HLA-DR (9). Song et al. (10) used single-cell RNA sequencing to compare transcriptional profiles of tumor and normal tissues from four patients with early-stage non-small cell lung cancer, and the authors confirmed the presence of cells with transcriptional features of MDSCs. In addition, immature (CD11b−CD13−/loCD16−), intermediate (CD11b+CD13−/loCD16−), and mature (CD11b+CD13+CD16+) neutrophils have been identified in the tumor microenvironment of patients with multiple myeloma (1, 11).
MDSCs inhibit immune responses through multiple immunosuppressive mediators, such as reactive nitrogen species and reactive oxygen species (ROS); depletion of amino acids, such as l-arginine and l-tryptophan; and secretion of cytokines, chemokines, and chemotactic factors (12). They express high levels of arginase-1 (Arg-1), NO synthase 2 (NOS2), NADPH oxidase 2 (NOX2), and cyclooxygenase-2 (COX2) in pathological settings, resulting in the production of Arg-1, NO, and ROS (13, 14). Arg-1 inhibits T cell proliferation by reducing local l-arginine (15), l-cysteine (16), and/or tryptophan levels (2, 17). NO reacts with multiple cellular compounds to generate regulatory and toxic factors. Finally, ROS, such as hydrogen peroxide (H2O2) and hydroxyl radical, induce direct oxidative damage in the nucleic acids, proteins, and lipids of the target cells (18).
MDSCs are generated from common myeloid progenitor cells in the bone marrow (BM) in response to multiple cues (2, 19), such as inflammatory cytokines, endoplasmic reticulum stress, metabolic stress, and oxidative/nitrosative and hypoxic stress. The differentiation and function of MDSCs are regulated by inflammation and tumor-dependent transcription factors, transcriptional coregulators, and chromatin-modifying factors (4, 9), along with cytokines such as GM-CSF and IL-6. The core transcription factors involved in the development of MDSCs include C/EBPβ, STAT3, and C/EBP homologous protein (CHOP) (1, 4, 9, 20, 21). Furthermore, noncoding RNAs, such as long noncoding RNAs (lncRNAs), can epigenetically regulate the genes involved in the differentiation and function of MDSCs through binding to chromatin-modifying factors and transcription factors (22). Multiple lncRNAs have been identified in myeloid immune cells (23–27). In this study, we detected a hitherto uncharacterized lncRNA, named lnc57Rik, in both human and murine MDSCs. Lnc57Rik was induced in the MDSCs by inflammation and tumor-associated factors. Furthermore, lnc57Rik mediated the immunosuppressive function of the MDSCs in vitro and in vivo by upregulating the Arg-1, NOS2, NOX2, and COX2 genes.
Materials and Methods
Mice, clinical tissue samples, and cell lines
Lnc57Rik (5730403I07Rik)-deficient mice on a C57BL/6J background were generated by the Model Animal Research Center of Nanjing University (Nanjing, Jiangsu, China) using the CRISPR-Cas9 system. C57BL/6 and B6.SJL-CD45a(Ly5a) (CD45.1) mice were also purchased from the Model Animal Research Center of Nanjing University (Nanjing, Jiangsu, China). OT-I and OT-II OVA-TCR transgenic mice were kindly provided by Dr. L. Lu of Zhejiang University. All mice were housed in specific pathogen-free conditions. The animal experiments were conducted in accordance with the Nankai University Guide for the Care and Use of Laboratory Animals. Peripheral blood and tissue samples were collected from patients with histologically confirmed colorectal adenocarcinoma at the People Union Hospital (Tianjin, China) after obtaining informed consent. The use of human tissue samples was approved by the Institute’s Human Ethics Committee of Nankai University and was in accordance with the Declaration of Helsinki. The murine melanoma cell line B16, human embryonic kidney cell line HEK 293T, and human leukemia monocytic cell line THP-1 were obtained from the American Type Culture Collection. B16-OVA cells were kindly provided by Dr. L. Chen from Yale University. The cell lines had not been authenticated in the past years and were confirmed to be free of mycoplasma contamination.
Recombinant murine or human GM-CSF and IL-6 were purchased from PeproTech (Rocky Hill, NJ). Inhibitors of STAT3 (HO-3867; Selleckchem), JAK1 (Filgotinib; GLPG0634, Selleckchem), JAK2 (JAK2-in-1; MCE), and STAT5 (STAT-in-2; MCE) were purchased as indicated. OVA MHC class II peptides (323–339) and OVA MHC class I peptides (257–264) were purchased from GenScript (Piscataway, NJ). Mouse anti-C/EBPβ and anti–histone H3 trimethylated lysine 4 (anti-H3K4me3) and rabbit anti-STAT3 and anti–homeotic 2-like protein (anti-ASH2L) Abs were purchased from Abcam (Cambridge, MA). Mouse anti-CHOP and anti-NOS2 Abs and rabbit anti-COX2 and anti–mixed lineage leukemia protein (anti-MLL) Abs were purchased from Cell Signaling Technology (Beverly, MA). Rabbit anti–Arg-1 and anti–gp91-phox Abs were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and mouse anti-V5 Ab was purchased from Thermo Fisher Scientific (Pittsburgh, PA). The anti–CD45-allophycocyanin, anti–CD45-PerCP/Cy5.5, anti–Gr1-FITC, anti–CD11b-PerCP/Cy5.5, anti–Ly6G-PE, anti–Ly6C-FITC, anti–CD4-FITC, anti–CD8a-PE, anti–CD8a-allophycocyanin, anti–IFN-γ-PE, anti–CD45.1-PE, and anti–CD45.2-allophycocyanin Abs were purchased from BD Biosciences (San Diego, CA).
In vitro induction of MDSCs and isolation of immune cells
For in vitro induction of MDSCs, BM cells (BMCs) were isolated from the femurs of C57BL/6 or lnc57Rik knockout (KO) mice and cultured in RPMI-1640 medium supplemented with GM-CSF (40 ng/ml) and/or IL-6 (40 ng/ml) for 4 d. The cells were harvested 0, 12, 24, 48, and 72 h later. To determine the dose-dependent effect of IL-6 or GM-CSF, we incubated the BMCs with different concentrations (10, 20, 40, and 80 ng/ml) of the respective cytokines to induce MDSC differentiation. Human MDSC-like cells were obtained by culturing peripheral blood monocytes with GM-CSF and IL-6 as previously described (24). Human PBMCs were isolated using a CD14+ isolation kit (R&D Systems) and then cultured in RPMI-1640 medium with 10% FCS and 1% penicillin/streptomycin in the presence of human recombinant GM-CSF (40 ng/ml) and IL-6 (40 ng/ml) for 4 d. CD11b+Gr1+ cells, CD4+ T cells, and CD8+ T cells were isolated by column-based magnetic separation using specific MACS Microbeads (Miltenyi Biotec) according to the manufacturer’s instructions.
Plasmid and small interfering RNAs transfection
Primary murine BMCs (C57BL/6 mice) and human peripheral blood monocytes were cultured in six-well plates and transfected with lnc57Rik small interfering RNA (siRNA), C/EBPβ liver-enriched activator protein (LAP) siRNA, WD repeat-containing protein 5 (WDR5) siRNA, negative control siRNA, or pcDNA3.1-lnc57Rik, pcDNA3.1-C/EBPβ LAP, pcDNA3.1-WDR5, and pcDNA3.1 control using Lipofectamine 3000 (Invitrogen) or HiPerFect transfection reagent (siRNA transfection; Qiagen) according to the manufacturer’s instructions. The transfected cells were then cultured in RPMI-1640 medium with 10% FBS (Life Technologies), 1% penicillin and streptomycin for 4 d in the presence of GM-CSF (40 ng/l) and IL-6 (40 ng/l) to induce MDSCs. All siRNAs were purchased from Sangon Biotech, and the sequences are listed in Supplemental Table I. The full-length sequences of C/EBPβ LAP/liver-enriched inhibitory protein (LIP)/WDR5 were amplified by PCR and directly cloned into the pcDNA3.1/V5-His TOPO TA plasmid (Invitrogen) using T4-conjugating enzymes. The PCR primer pairs are listed in Supplemental Table I. As per the experimental requirements, the suitably transfected MDSCs were incubated with 20 nM GLPG0635, 100 nM HO-3867, 150 nM JAK2-in-1, or 100 nM STAT-in-2 for 24 h.
The suitably treated cells were harvested and rinsed twice with ice-cold PBS and then incubated with FITC-, PE-, PerCP-, or allophycocyanin-labeled Abs for 30 min in PBS with 1% FBS as described previously (28). After washing twice, the cells were resuspended in PBS and analyzed in a FACScan flow cytometer (BD Biosciences). For T cell proliferation, MDSCs were cocultured with CFSE-labeled CD3+ T cells at 1:1 ratio in the presence of anti-CD3/anti-28 Dynabeads for 72 h. T cell proliferation was analyzed by flow cytometry. Dead cells were eliminated through Zombie or 7-aminoactinomycin D staining.
Expression levels of lncRNAs, microRNAs, and coding mRNAs in wild type (WT) and lnc57Rik KO mice were analyzed by Beijing Capitalbio Technology. In brief, total RNA was extracted using TRIzol (Life Technologies), and the contaminating DNAs were removed using RNeasy spin columns (Qiagen). The quality of isolated RNA samples was evaluated with an Agilent Bioanalyzer 2100 (Agilent Technologies), and the purified RNA was quantified using a NanoDrop ND-2000 spectrophotometer (Infinigen Biotech). The Agilent Gene Expression oligo microarrays and miRNA microarrays were analyzed using the R software platform, and the limma (a linear regression model) package was used to screen for the differentially expressed genes with fold change > 2 or < −2 and an adjusted p < 0.05 as the thresholds.
Establishment of s.c. xenografts and BM transplantation
The WT and lnc57Rik KO mice were s.c. injected with 1 × 106 OVA-B16 cells at the inguinal site. For the BM transplantation experiments, CD45.1 recipient mice were injected with the melanoma cells as earlier, irradiated with a single dose of 800 cGy using a Shepherd Mark I Cesium Irradiator (J.L. Shepherd and Associates), and then inoculated with the BMCs isolated from WT or lnc57Rik KO mice (1 × 106 cells/mouse) via the tail vein. The length (a) and width (b) of the s.c. tumors were measured using Vernier’s caliper every 2 d, and the tumor volume was calculated as ab2/2. Twenty-two days postinoculation, the tumors were dissected and homogenized into single-cell suspensions. The tumor cells were stained using the requisite Abs, and the proportions of CD4+ T cells, CD8+ T cells, CD4+IFN-γ+ cells, and CD8+IFN-γ+ cells were analyzed by flow cytometry.
Establishment of azoxymethane/dextran sodium sulfate–induced model of colitis-associated carcinogenesis
WT and lnc57Rik KO mice were subjected to dextran sodium sulfate (DSS)/azoxymethane (AOM)-induced colon carcinogenesis as previously reported (29). In brief, 7- to 8-wk-old mice were injected i.p. with 12.5 mg/kg AOM (Sigma) dissolved in 0.9% NaCl. At 1 wk after injection, the mice were given drinking water supplemented with 2% DSS, followed by regular water for 2 wk. Two weeks after the third cycle, the mice were sacrificed and their colons were removed. After flushing out the contents with PBS, the colons were cut longitudinally and fixed flat in 10% neutral-buffered formalin overnight. The tumor foci were counted and measured under a stereomicroscope. Thereafter, the tissues were embedded in paraffin for H&E staining or embedded in OCT compound and frozen in liquid nitrogen for immunostaining.
Lnc57Rik deletion in THP-1 cells
As reported in our previous study (30), multiple target sites were selected for the lnc57Rik gene, and one small guide RNA was designed for each target site. Oligonucleotides were synthesized according to the small guide RNA sequences. The Cas9 plasmid containing the GFP tag and screening gene (puromycin) was linearized and ligated with the oligonucleotide using T4 ligase in an overnight reaction at 16°C. The verified plasmid was amplified in bacterial cells and extracted as per standard protocols. THP-1 cells were seeded in 12-well plates at a density of 100,000 cells/well and transiently transfected 24 h later with 4 μg of Cas9 plasmid using Lipofectamine 2000 (Life Technologies). After culturing the cells for 24 h, puromycin was added to the medium at a final concentration of 3 µg/µl. After 3 d of selection, the stably transfected cells were harvested and resuspended in fresh medium at a density of 1–2 cells/200 µl. The transfected cells were seeded in a 96-well plate, and the number of wells containing single cells was marked after 5 d of culture. Once the cells in the marked wells were ∼70–90% confluent, they were harvested and the DNA was extracted and amplified with the previously designed PCR primers. The amplified products are sequenced to verify lnc57Rik deletion.
The immunosuppressive function of MDSCs was evaluated through biochemical assays as previously reported (31). To measure arginase activity, we lysed the cells with 100 μl of 0.1% Triton X-100 at 4°C for 30 min. The lysate was mixed with 100 μl of 25 mM Tris–HCl and 10 μl of 10 mM MnCl2, heated at 56°C for 10 min, and then incubated with 100 μl of 0.5 M l-arginine (pH 9.7) at 37°C for 120 min. The reaction was stopped with 900 μl of H2SO4 (96%)/H3PO4 (85%)/H2O (1:3:7). After adding 40 μl of 9% α-isonitrosopropiophenone, the solution was heated at 95°C for 30 min, and the urea concentration was determined by measuring the absorbance at 540 nm. The Nitrate/Nitrite Assay Kit (Kamiya) was used to measure NO production. In brief, certain volumes of 2 mM NADPH (5 μl), 2 mM favin adenine dinucleotide (10 μl), and 2 mM nitrate reductase (5 μl) were added to the cell lysate (60 μl), and the mixture was incubated at 37°C for 30 min, followed by addition of 10 μl of lactate dehydrogenase buffer and incubation for another 30 min at 37°C. Then, 50 μl of Griess Reagent I and Griess Reagent II was added, and the solution was further incubated at room temperature for 10 min. Next, the absorbance at 540 nm was measured. H2O2 production was evaluated using the Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen). In brief, 1 × 104 cells were resuspended in Krebs–Ringer phosphate buffer containing 50 mM Amplex Red reagent and 0.1 U/ml HRP. After adding PMA to a final concentration of 30 ng/ml, the absorbance at 560 nm was measured using a microplate reader at 37°C. ROS production was measured using the oxidation-sensitive dye dichlorodihydrofluorescein diacetate (DCFDA). The cells were incubated in RPMI medium with 2.5 μM DCFDA at 37°C for 30 min. For PMA-induced activation, cells were simultaneously cultured with DCFDA and 30 ng/ml PMA and then analyzed by flow cytometry.
In vitro evaluation of MDSC-mediated immunosuppression
Splenic cells obtained from OT-I or OT-II mice were seeded in 96-well plates at a density of 2 × 106 cells/well and stimulated with 200 nM MHC class I OVA peptide (OVA257–264, 1 μg/ml) or MHC class II OVA peptide (OVA323–339, 1 μg/ml). The MDSCs from WT or lnc57Rik KO mice were then cocultured with the OT-I or OT-II splenic cells at different ratios (2 × 106 cells/well as the starting density) in RPMI-1640 medium for 48 h. IFN-γ levels in the supernatant were measured using an ELISA kit (Biotech) according to the manufacturer’s instructions. For the negative control, 2 × 106 OT-I or OT-II cells were cultured alone. For the positive control, 2 × 106 OT-I or OT-II cells were cultured with 1 μg/ml peptide.
Chromatin immunoprecipitation (ChIP)-PCR was performed using EZ-ChIP Chromatin Immunoprecipitation Kit (Millipore) as previously reported (30, 32). In brief, the cells were cross-linked with 1% paraformaldehyde and incubated at room temperature with constant shaking. The cross-linking reaction was stopped after 10 min by adding glycine to a final concentration of 0.125 M for 5-min incubation. The cross-linked cells were washed thrice with ice-cold PBS (containing 1% PMSF) and immediately resuspended in SDS lysis buffer (containing 1% PMSF). Cell lysates were sonicated on ice for 40 cycles of 30 s ON and 30 s OFF in 10-cycle increments using a Bioruptor (Diagenode). After pelleting the debris, the lysates were cleared by incubating with protein G-agarose for 1 h at 4°C with rotation. For immunoprecipitation (IP), the precleared cell lysate was incubated overnight with the indicated Abs at 4°C with rotation, and protein G-agarose was added for the final 2 h of incubation. The beads were washed with low-salt and high-salt LiCl wash buffer, and the chromatin immunocomplex was eluted using elution buffer at room temperature for 15 min. DNA was removed from the protein–DNA complexes by incubating overnight with 5 M NaCl at 65°C and then treated with RNase at 37°C for 30 min and Proteinase K at 55°C for 2 h. Finally, the purified DNA was amplified by quantitative PCR (qPCR).
Immunostaining and RNA fluorescence in situ hybridization
Immunostaining and RNA fluorescence in situ hybridization (RNA-FISH) were performed according to our previously reported protocol (33). In brief, cells were first slicked on sterile and 0.01% poly-lysine–treated slides in a six-well plate. The slides were then sequentially treated with ice-cold cytoskeletal (CSK) buffer, CSK+0.4% Triton X-100 buffer, and CSK buffer for 30 s each to permeabilize the cell membrane. After fixing the cells with 4% PFA for 10 min and then with cold 70% ethanol three times, the slides were rinsed thrice with ice-cold PBS and then incubated in prewarmed 5% goat serum for 30 min at 37°C to block nonspecific binding. The slides were incubated with the primary Ab at 37°C for 1 h, washed thrice with 1× PBS/0.2% Tween 20 for 3 min on a rocker, and then incubated with the secondary Ab at 37°C for 30 min. After dehydrating through an ethanol gradient (85, 95 and 100% for 2 min each) at room temperature, the slides were air-dried at room temperature for 15 min and hybridized overnight with the indicated probes at 37°C in a humid chamber. After hybridization, the slides were washed with 2× SSC/50% formamide, 2× SSC, and 1× SSC for three times each and counterstained with DAPI. Finally, the slides were sealed and observed under a confocal microscope.
RNA immunoprecipitation (RIP) was performed as previously reported (30, 32). In brief, the cells were harvested, washed, and lysed with ice-cold IP lysis buffer (Thermo Scientific Pierce) containing 0.5% RNase inhibitor (Invitrogen). The lysates were incubated on ice for 5 min with periodic mixing and then centrifuged at 13,000 × g for 10 min at 4°C. The supernatants were transferred into a new tube and incubated with protein G-agarose for 1 h at 4°C with rotation. The precleared lysate was then incubated overnight with the suitable Ab at 4°C with rotation. Protein G-agarose was removed by centrifuging at 3000 × g for 1 min, and the immunoprecipitate was washed sequentially with IP lysis buffer (containing 0.5% RNase inhibitor). Finally, RNA was extracted from the protein–RNA complexes on the beads using TRIzol reagent, dissolved in diethyl pyrocarbonate water, and quantified by qPCR.
RNA–protein pull-down assay was performed using the Pierce Magnetic RNA-Protein Pull-Down Kit as reported previously (30, 32). The suitably treated cells were harvested and lysed in the IP lysis buffer (Thermo Scientific Pierce). Lnc57Rik was transcribed in vitro (manual HiScribe T7 in vitro transcription Kit; NEB) and labeled using RNA 3′ Desthiobiotinylation Kit (Thermo Scientific Pierce). The labeled lnc57Rik was bound to streptavidin magnetic beads by incubating with 50 μl of beads and 50 pmol-labeled RNA in the RNA capture buffer for 30 min at room temperature with constant agitation. The beads were washed with an equal volume of 20 mM Tris (pH 7.5), mixed with 100 μl of protein–RNA binding buffer, and then incubated with 100 μl of RNA–protein reaction mix for 60 min at 4°C with rotation. After washing the beads twice with 100 µl of wash buffer, 50 μl of elution buffer was added and the reaction mixture was incubated for 30 min at 37°C with agitation. The samples were analyzed using a gel.
Statistical analyses were performed using two-tailed Student t test and one-way ANOVA Bonferroni’s multiple comparison test. Survival curves were plotted by the Kaplan–Meier method and compared using the generalized Wilcoxon test. A p value <0.05 with a 95% confidence interval was considered statistically significant.
Tumor-associated factors promote the expression of lnc57Rik in MDSCs
lncRNAs play an important role in regulating the differentiation and function of MDSCs (33–35). To this end, we analyzed the transcriptomes of MDSCs exposed to the tumor-associated factor IL-6, a major regulator of MDSCs (36), and the untreated control cells through microarrays. As shown in (Fig. 1A, there were significant differences in the expression levels of mRNAs and lncRNAs between the two groups (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104718).
Several upregulated lncRNAs, including 5730403I07Rik, were further confirmed in the MDSCs (Supplemental Fig. 1A). However, except the reported lncRNA Olfr29-ps-1 (33), lncRNA 5730403I07Rik, also known as lnc57Rik, could regulate the suppressive function of MDSCs (Supplemental Fig. 1B, 1C). The expression of lnc57Rik in MDSCs could also be confirmed using Northern blot and FISH (Fig. 1B, 1C). The MDSCs treated with IL-6 or GM-CSF expressed high levels of lnc57Rik, which was confirmed by qRT-PCR and FISH (Fig. 1D, 1E). Moreover, IL-6 or GM-CSF upregulated lnc57Rik in the MDSCs in a time- and dose-dependent manner, with peak levels at 24 h at the concentration of 40 ng/ml (Fig. 1F–H). Because IL-6 can activate the JAK1/STAT3 pathway (37), we also assessed the effect of STAT3 and JAK1 inhibitors on lnc57Rik expression. As expected, lnc57Rik was markedly downregulated in the MDSCs treated with the specific inhibitors (Fig. 1I). Furthermore, inhibitors of JAK2 and STAT5, the downstream effectors of GM-CSF, also decreased lnc57Rik expression levels in the MDSCs (Fig. 1I). Consistent with the findings so far, the MDSCs isolated from tumor tissues expressed significantly higher levels of lnc57Rik compared with the BM-derived MDSCs (Fig. 1J, 1K). In addition, lnc57Rik was expressed only in myeloid-derived cells such as MDSCs, macrophages, and dendritic cells (DCs), but not in the CD4+ and CD8+ T cells and B cells (Fig. 1L). These data collectively indicate that tumor-associated factors such as IL-6 and GM-CSF can induce lnc57Rik expression in the resident MDSCs.
Lnc57Rik positively regulates the immunosuppressive function of MDSCs in vitro
In vitro gain- and loss-of-function studies were performed to determine whether lnc57Rik regulates MDSC-mediated immune suppression (Fig. 2A). Arg-1, CYBB (NOX2), NOS2, and ptgs2 (COX2), the critical genes involved in MDSC-mediated immunosuppression (13, 14), were significantly downregulated in the lnc57Rik knockdown MDSCs and upregulated in the MDSCs overexpressing exogenous lnc57Rik (Fig. 2B, 2C). Not surprisingly, lnc57Rik knockdown and overexpression also had similar effects on the products of the earlier genes, i.e., Arg-1, NO, H2O2, and ROS (Fig. 2D–G). Because these metabolites influence the proliferation and function of CD4+ and CD8+ T cells (38, 39), we next investigated the effects of MDSCs on the proliferation of OVA-specific OT-I CD8+ or OT-II CD4+ splenic cells and measured IFN-γ release. As shown in (Fig. 2H, lnc57Rik knockdown markedly reduced the suppressive function of MDSCs on T cell proliferation, whereas exogenous lnc57Rik promoted MDSC-mediated suppression. Furthermore, the CD4+ and CD8+ T cells cocultured with lnc57Rik-knockdown MDSCs produced higher levels of IFN-γ in the presence of OVA MHC class I or class II peptides compared with those cocultured with the control MDSCs. In contrast, IFN-γ production was markedly reduced in the T cells cocultured with MDSCs overexpressing exogenous lnc57Rik (Fig. 2I). These results clearly indicated that lnc57Rik is a key driver of the immunosuppressive function of MDSCs. To further validate this hypothesis, we generated lnc57Rik KO mice and performed a microarray analysis of the MDSCs isolated from lnc57Rik KO and WT mice. As expected, Arg-1, NOX2, and NOS2 were downregulated in the lnc57Rik KO MDSCs compared with the WT cells (Fig. 2J) (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE197566), and the results were confirmed by qRT-PCR and immunoblotting (Fig. 2K, 2L). Likewise, Arg-1 activity and the production of NO, H2O2, and ROS were markedly diminished in the lnc57Rik KO MDSCs (Fig. 2M–P). Furthermore, the MDSCs from lnc57Ri KO mice had a significantly weaker inhibitory effect on T cell proliferation compared with the controls (Fig. 2Q). These MDSCs from lnc57Ri KO mice also had reduced suppression on the CD4+ and CD8+ T cells as compared with the MDSCs from WT mice (Fig. 2R). In addition, the percentages of the CD11b+Ly6G+Ly6C+ and CD11b+Ly6G−Ly6C+ MDSC subsets also differed between the lnc57Rik KO and WT groups (Fig. 2S), indicating that lnc57Rik likely affects the differentiation of MDSCs. Taken together, lnc57Rik promotes the immunosuppressive function of MDSCs.
Lnc57Rik promotes the immunosuppressive function of MDSCs in vivo
To determine whether lnc57Rik regulates the immunosuppressive function of MDSCs in the tumor microenvironment, we established s.c. OVA-expressing B16 melanoma xenografts in WT and lnc57Rik KO mice. The tumors in the lnc57Rik KO mice were significantly smaller and lighter compared with those in WT mice (Fig. 3A–C). The slower tumor growth in the lnc57Rik KO mice was accompanied by a significantly increased infiltration of CD4+ and CD8+ T cells in the tumor tissues (Fig. 3D), along with higher in situ levels of IFN-γ (Fig. 3E). Consistent with the earlier results, fewer DSS/AOM-induced colonic tumors and more IFN-γ–producing immune cells were observed in the lnc57Rik KO mice compared with the WT controls (Fig. 3F–M). These findings suggested that lnc57Rik can promote tumor growth by promoting an immunosuppressive microenvironment, although the role of other BM-derived immune cells or stromal cells could not be ruled out. Therefore, to confirm that the MDSCs were the key immunosuppressive cells involved in promoting tumor growth, we generated chimeric mice by transplanting BMCs from the lnc57Rik KO or WT mice into the WT tumor-bearing recipients after sublethal whole-body irradiation (Fig. 3N). Transplantation of lnc57Rik KO BMCs markedly suppressed tumor growth in the recipient mice, whereas the WT MDSCs had no significant effect (Fig. 3O–Q), suggesting that tumor suppression indeed depends on the BM-derived MDSCs. The proportion of CD4+ and CD8+ T cells (Fig. 3R) and that of CD4+IFN-γ+ and CD8+IFN-γ+ cells (Fig. 3S) also increased significantly in the tumor tissues of mice injected with the lnc57Rik KO BMCs than those with WT BMCs. Furthermore, MDSCs from the tumor tissues of lnc57Rik KO mice bearing OVA-B16 tumors also exhibited a weaker suppressive effect on T cell proliferation in vitro (Fig. 3T). Consistent with this, the CD4+ and CD8+ T cells isolated from the tumor tissues of mice injected with the lnc57Rik KO BMCs produced higher levels of IFN-γ after exposure to MHC class II– or class I–restricted OVA peptides, which was also indicative of the weaker suppressive effects of lnc57Rik KO MDSCs on immune effector cells (Fig. 3U). CD45.2+ cells were detected in the spleen and tumor tissues of recipient mice (Fig. 3V, 3W). Taken together, the induction of lnc57Rik in the tumor-associated MDSCs can improve the latter’s immunosuppressive function and promote tumor growth.
Lnc57Rik promotes transcription of immunosuppressive genes on binding with the C/EBPβ isoform LAP and activating C/EBPβ
lncRNAs regulate gene expression by interacting with miRNAs, mRNAs, and proteins and by encoding small peptides (40, 41). Previous studies have shown that lncRNAs also recruit and bind to transcription factors and epigenetic regulators (42–44). We identified the putative binding proteins of lnc57Rik through bioinformatics analysis (Supplemental Fig. 2A). Because multiple transcription factors, such as C/EBPβ, STAT3, and CHOP, play a central role in the differentiation and function of MDSCs (21, 45, 46), we hypothesized that lnc57Rik might combine with these factors to regulate MDSC function. To this end, we performed RIP using anti-C/EBPβ, anti-STAT3, or anti-CHOP Abs and found that only C/EBPβ bound with lnc57Rik (Fig. 4A, Supplemental Fig. 2B, 2C). This finding was further confirmed by FISH (Fig. 4B). There are three isoforms of C/EBPβ, namely, LPA*, LAP, and LIP, which are translated from unique C/EBPβ transcripts from different initiation sites (47, 48). To identify the C/EBPβ isoform that bound to lnc57Rik, we generated V5-tagged C/EBPβ-LAP and C/EBPβ-LIP sequences (Fig. 4C). The pull-down assay showed that lnc57Rik bound with the LAP isoform of C/EBPβ (Fig. 4D) in a dose-dependent manner (Fig. 4E). Furthermore, the promoter regions of the lnc57Rik-regulated genes (Arg-1, NOS2, COX2, and NOX2) also harbored potential binding sites for C/EBPβ LAP (Fig. 4F). This was confirmed by ChIP-PCR, which showed that lnc57Rik KO reduced the enrichment of C/EBPβ in these promoter regions (Fig. 4G). C/EBPβ-LAP knockdown and exogenous LAP expression further indicated that the binding of lnc57Rik with C/EBPβ-LAP is necessary for the regulatory effect of lnc57Rik on the expression of the aforementioned genes (Fig. 4H–O). Taken together, lnc57Rik can interact with the C/EBPβ isoform LAP to induce immunosuppressive genes in the MDSCs.
Lnc57Rik promotes transcription of immunosuppressive genes through H3K4me3 methylation and enrichment
H3K4me3 promotes gene transcription (49, 50), and several lncRNAs have been identified that can enrich H3K4me3 marks on the target genes (51–53). Therefore, we also examined H3K4me3 marks in the promoter regions of Arg-1, NOS2, COX2, and NOX2. ChIP-PCR demonstrated that lnc57Rik KO reduced the enrichment of the H3K4 me3 marks in the promoter region of these genes, whereas lnc57Rik had the opposite effect (Fig. 5A). Thus, lnc57Rik may increase the expression levels of these genes in the MDSCs by enriching the H3K4me3 marks in their promoter regions. H3K4me3 methylation and enrichment depend on a core methyltransferase complex consisting of WDR5, ASH2L, MLL1, and retinoblastoma-binding protein 5 (53, 54). Consistent with this, RIP experiments showed that lnc57Rik could also bind to WDR5, ASH2L, and MLL (Fig. 5B, Supplemental Fig. 3A, 3B). The results were confirmed by RNA-FISH and pull-down analysis (Fig. 5C, 5D). To further determine whether WDR5 mediates the functions of lnc57Rik in MDSCs, we knocked down and overexpressed WDR5 in the MDSCs. Data demonstrated that the binding of lnc57Rik with WDR5 was also necessary for the expression of Arg-1, NOS2, COX2, and NOX2 genes (Fig. 5E–L). Taken together, lnc57Rik induced immunosuppressive genes in the MDSCs by enriching H3K4me3 marks in their promoters via WDR5.
Human lnc57Rik has similar functions as murine lnc57Rik
Bioinformatics analysis showed high sequence homology between human and murine lnc57Rik (http://blast.ncbi.nlm.nih.gov/Blast.cgi), which led us to hypothesize that the human lnc57Rik may have a similar function as its murine counterpart. Human MDSC-like cells were obtained by culturing human peripheral blood monocytes with GM-CSF and IL-6 (Fig. 6A, 6B). Microarray analysis showed significant differences in the transcriptomic profiles of the ln57Rik-knockdown and overexpressing human MDSCs compared with their respective controls (Fig. 6C, 6D) (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE197567). Knocking down lnc57Rik in the MDSCs downregulated Arg-1, NOS2, NOX2, and COX2, whereas exogenous lnc57Rik overexpression led to the upregulation of these genes (Fig. 6E–H). Furthermore, MDSCs transfected with human lnc57Rik siRNA showed a weaker suppressive effect on T cell proliferation, while overexpression of lnc57Rik in these cells increased immunosuppression (Fig. 6I). To further confirm this result, we deleted human lnc57Rik in the THP-1 monocyte cell line using the CRISPR/Cas9 technique. Arg-1, NOS2, NOX2, and COX2 were significantly downregulated in the lnc57Rik KO THP1 cells (Fig. 6J, 6K). Consistent with this, Arg-1 activity and the production of NO and H2O2 were also reduced in the lnc57Rik KO THP-1 cells compared with the controls (Fig. 6L–N). Both RIP and RNA-FISH analyses showed that human lnc57Rik could bind to C/EBPβ and WDR5 in human MDSCs (Fig. 6O–R). Finally, colon tumor specimens resected from patients had a higher proportion of lnc57Rik-overexpressing MDSCs compared with the peripheral normal tissues (Fig. 6S). Taken together, human lnc57Rik has a similar function as murine lnc57Rik.
In this study, we found that lnc57Rik is upregulated in the MDSCs by IL-6 and promotes their immunosuppressive function. It binds to the LAP subunit of C/EBPβ and the H3K4me3 methyltransferase WDR5, which transcriptionally activates the genes involved in MDSC-mediated immunosuppression, including Arg-1, NOS2, COX2, and NOX2. The growth of s.c. melanoma xenografts and DSS-induced colonic tumors was markedly inhibited in the lnc57Rik KO mice compared with the WT mice, which further confirmed that lnc57Rik mediates the immunosuppressive function of MDSCs and promotes tumor growth. Thus, lnc57Rik can be potentially targeted to control the differentiation and suppressive function of MDSCs in inflammatory diseases and cancer.
The binding of lnc57Rik with LAP and WDR5 transcriptionally activated the genes involved in the immunosuppressive function of MDSCs, i.e., Arg-1, NOS2, COX2, and NOX2. C/EBPβ, a transcription factor belonging to the leucine zipper family, is critical to the immunosuppressive program in both tumor-induced and BM-derived MDSCs (55). Three isoforms of C/EBPβ have been identified, including LAP*, LAP, and LIP, which differ in their physiological roles. While LAP* participates in terminal differentiation, LAP and LIP promote cell proliferation and tumor progression. LAP* and LAP can promote differentiation of the granulocytic lineages and induce the expression of immunosuppressive genes (20, 56). Furthermore, C/EBPβ LAP can heterodimerize with other family members such as LIP to inhibit the transcriptional activity of LAP (57). We found that lnc57Rik binds with LAP to promote the expression of the aforementioned immunosuppressive genes, suggesting that binding of lnc57Rik with C/EBPβ-LAP frees LAP from the inhibitory C/EBPβ LAP/LIP complex. Previous studies have shown that C/EBPβ plays a central role in regulating the immunosuppressive function of MDSCs (9, 55). The DNA methyltransferase WDR5 regulates multiple biological processes, including the self-renewal and reprogramming of embryonic stem cells (58) and the function of immune cells through histone H3K4 trimethylation (59, 60). WDR5 epigenetically activates gene expression levels by enriching H3K4me3 marks on the promoters. Furthermore, it is known to bind with the lncRNA HOTTIP and promote transcription of the target genes through H3K4 trimethylation (61). Knocking down or knocking out WDR5 can reduce the expression of genes regulated by its target lncRNAs (62). Our data indicate that WDR5 also binds with lnc57Rik in the MDSCs to regulate the expressions of the downstream immunosuppressive genes.
Altogether, our data show that lnc57Rik is robustly induced in inflammatory and tumor microenvironments and promotes the differentiation of the resident MDSCs and their immunosuppressive effects. Although we did not investigate the effect of lnc57Rik on DCs, we cannot rule out the possibility that it may also regulate the immune function of DCs. Two subpopulations of MDSCs have been identified, namely, the mononuclear-MDSCs and polymorphonuclear-MDSCs. It remains to be elucidated whether lnc57Rik affects these two subsets differentially. In addition, in the BM of chimera mice, other cell populations such as DCs may also exert a role in the development of diseases.
This work was supported by the National Natural Science Foundation of China (91842302, 81970488, 81970457, and 91629102), Tianjin Science and Technology Commission (20JCQNJC01780 and 18JCZDJC35300), Ministry of Science and Technology (2016YFC1303604), and the State Key Laboratory of Medicinal Chemical Biology and the Fundamental Research Funds for the Central Universities, Nankai University (63191724).
The online version of this article contains supplemental material.
The sequences presented in this article have been submitted to the Gene Expression Omnibus under accession numbers GSE197566 and GSE197567.
Abbreviations used in this article:
homeotic 2-like protein
bone marrow cell
C/EBP homologous protein
- CSK, cytoskeletal; DC
- DCFDA, dichlorodihydrofluorescein diacetate; DSS
dextran sodium sulfate
fluorescence in situ hybridization
histone H3 trimethylated lysine 4
long noncoding RNA
- LAP, liver-enriched activator protein; MDSC
myeloid-derived suppressor cell
mixed lineage leukemia protein
NO synthase 2
NADPH oxidase 2
polymorphonuclear myeloid-derived suppressor cell
quantitative real-time PCR
reactive oxygen species
small interfering RNA
WD repeat-containing protein 5
The authors have no financial conflicts of interest.