The miR-23a∼27a∼24-2 microRNA Cluster Promotes Inflammatory Polarization of Macrophages

Austin Boucher, Nathan Klopfenstein, William Morgan Hallas, Jennifer Skibbe, Andrew Appert, Seok Hee Jang, Kirthi Pulakanti, Sridhar Rao, Karen D. Cowden Dahl and Richard Dahl
J Immunol February 1, 2021, 206 (3) 540-553; DOI: https://doi.org/10.4049/jimmunol.1901277

Key Points

  • mirn23a miRNAs repress inflammatory signaling in myeloid progenitors.

  • mirn23a−/− macrophages have enhanced M2 polarization and decreased responses to LPS.

  • mirn23a−/− macrophages enhance the growth of syngeneic tumors in vivo.

Abstract

Macrophages are critical for regulating inflammatory responses. Environmental signals polarize macrophages to either a proinflammatory (M1) state or an anti-inflammatory (M2) state. We observed that the microRNA (miRNA) cluster mirn23a, coding for miRs-23a, -27a, and -24-2, regulates mouse macrophage polarization. Gene expression analysis of mirn23a-deficient myeloid progenitors revealed a decrease in TLR and IFN signaling. Mirn23a−/− bone marrow–derived macrophages (BMDMs) have an attenuated response to LPS, demonstrating an anti-inflammatory phenotype in mature cells. In vitro, mirn23a−/− BMDMs have decreased M1 responses and an enhanced M2 responses. Overexpression of mirn23a has the opposite effect, enhancing M1 and inhibiting M2 gene expression. Interestingly, expression of mirn23a miRNAs goes down with inflammatory stimulation and up with anti-inflammatory stimulation, suggesting that its regulation prevents locking macrophages into polarized states. M2 polarization of tumor-associated macrophages (TAMs) correlates with poor outcome for many tumors, so to determine if there was a functional consequence of mirn23a loss modulating immune cell polarization, we assayed syngeneic tumor growth in wild-type and mirn23a−/− mice. Consistent with the increased anti-inflammatory/immunosuppressive phenotype in vitro, mirn23a−/− mice inoculated with syngeneic tumor cells had worse outcomes compared with wild-type mice. Coinjecting tumor cells with mirn23a−/− BMDMs into wild-type mice phenocopied tumor growth in mirn23a−/− mice, supporting a critical role for mirn23a miRNAs in macrophage-mediated tumor immunity. Our data demonstrate that mirn23a regulates M1/M2 polarization and suggests that manipulation of mirn23a miRNA can be used to direct macrophage polarization to drive a desired immune response.

Introduction

Macrophages are critical effector cells of the innate immune system that respond to danger signals from pathogens and tissue damage (1). Present in every tissue, they are first responders to challenges from the environment. In the absence of host challenge, macrophages will act to promote tissue repair and development, often clearing away dead cells (2). Upon pathogen challenge, macrophages switch their function to cause local, nonspecific inflammation in an attempt to clear the challenge and recruit other immune cells to help. Upon clearance of the pathogen, macrophages will return to their previous functional state and repair damage generated by the inflammatory response (3).

Macrophage plasticity is an important factor for both promoting and resolving inflammation (4). Classically activated M1 proinflammatory macrophages are generated in response to TLR ligands such as LPS and IFN-γ, whereas alternatively activated M2 anti-inflammatory macrophages are generated in response to Th2 cytokines such as IL-4 and IL-13. The ability to toggle between these two activated states along the M1/M2 spectrum makes macrophages a crucial part of immune responses in multiple disease environments. Thus, re-education of macrophages in disease contexts could act as a therapeutic strategy for treatment.

Molecular regulation of macrophage polarization is complex. TLR ligands, TNF, and IFNs promote M1 polarization, whereas M-CSF, IL-4, IL-10, and IL-13 (alone and/or in combination) promote acquisition of an M2 phenotype (5). Transcription factors are activated downstream of these soluble signals and include M1-promoting factors NF-κB, AP-1, IRF1, IRF5, and STAT1 and M2-promoting factors STAT3, -5, and -6, IRF4, and PPARγ (6, 7). Epigenetic modification such as DNA and histone methylation as well as histone acetylation influence macrophage polarity through regulating the accessibility of transcription factors to chromatin (8, 9). Recently, microRNAs (miRNAs) were shown to be an additional layer of genetic regulation in macrophage polarization (10). miRNAs are small, noncoding RNAs that target the 3′ untranslated region of their target mRNAs that result in degradation/translational repression of their target transcripts (11). miRNA profiling demonstrates that over 100 miRNAs have differential expression in murine bone marrow–derived macrophages (BMDMs) between M1- and M2-polarizing conditions (12). miRNAs upregulated in M1-polarized macrophages compared with M2 macrophages include miR-155-5p, miR-181a, and miR-451. miRNAs downregulated in M1 cells compared with M2 include miR-125-5p, miR-143-3, miR-145-5p, and miR-146a-3p. Altering miRNA expression impacts macrophage polarization. Mir155−/− mice are refractory to the development of colitis induced by inflammatory chemical stimuli (13). This protection is mediated by increased M2 polarization of the mir155−/− macrophages. Macrophages treated with microvesicles carrying miR-155 induced an M1 phenotype (14). MiR-146a has the opposite effect on macrophage polarization. Mir146a−/− macrophages are hyperresponsive to LPS stimulation with increased production of proinflammatory cytokines TNF, IL-6, and IL-1β (15). Conversely, overexpression of miR-146a in peritoneal macrophages enhances M2 polarization and decreases M1 polarization (16). Inflammation requires precise regulation and current data suggest that miRNAs are involved in buffering inflammatory signaling. The ability of miRNAs to buffer the expression of multiple transcription factors and signaling transducers that regulate macrophage polarization make them attractive targets for manipulating immune responses.

Previously, we demonstrated that the miRNA gene cluster mirn23a, coding for three mature miRNAs miRs-23a, -27a, and 24-2, promotes myeloid development at the expense of B cell development using overexpression and genetic knockout models (17, 18). In these studies, we showed that mirn23a targets genes crucial to cell fate decisions in hematopoietic stem and progenitor cells (HSPCs) (19). The mirn23a miRNAs are expressed in HSPCs and continue to be expressed in mature hematopoietic cells with highest expression observed in monocytes and granulocytes (17).

Mirn23a miRNAs are implicated in the regulation of macrophage function, but it is not clear from overexpression and inhibitor studies whether these miRNAs promote or repress inflammatory phenotypes. Importantly, the majority of studies have examined individual miRNAs of the cluster and not studied the effect of simultaneously overexpressing or antagonizing all three cluster miRNAs. MiR-27a is upregulated in both human and mouse macrophages by LPS treatment, and 3′ untranslated region luciferase assays demonstrate that miR-27a has the ability to directly repress the expression of the anti-inflammatory cytokine IL-10 (20). Overexpression of miR-27a mimics decreased production of IL-10 in macrophages, whereas miR-27a inhibitors have the opposite effect (20, 21). In addition to inflammatory effects mediated through the targeting of IL-10, miR-27a was also observed to augment TLR-mediated production of type I IFNs through the targeting of Siglec1 and Trim27 (22). However, other reports suggest that miR-27a induces M2 anti-inflammatory polarization of macrophages (23, 24). Two reports suggest that miR-24 promotes M2 anti-inflammatory polarization of macrophages. MiR-24 overexpression was shown to antagonize expression of M1-related molecules TNF, IL-6, IL-12, and inducible NO synthase (iNOS) and conversely promote the expression of M2-related molecules CD206 and Arg-1 (25, 26). However, contradicting results were obtained with miR-24 mimic expression in RAW264.7 murine macrophages, which resulted in decreased expression of M2 genes Arg1, Fizz1, and Il10 (27). The same group showed that miR-24 inhibitors decreased expression of these M2 markers. Fewer studies have addressed the role of miR-23a in macrophage function; however, two studies both supported a role for promoting inflammatory polarization of macrophages (28, 29).

So far, all studies examining mirn23a miRNA function in macrophages have only manipulated expression of one miRNA at a time and have used overexpression and knockdown approaches. To address the discrepancies in these models of nonphysiological expression, we used mirn23a germline knockout mice to elucidate the function of mirn23a in macrophage polarization. We observe that mirn23a miRNAs regulate macrophage responses associated with polarization. The majority of mirn23a’s affects appear to be proinflammatory. RNA sequencing (RNA-seq) data reveal a strong downregulation of inflammatory signaling pathways in mirn23a−/− myeloid progenitors. In mature mirn23a-deficient macrophages, we observe increases in production of the anti-inflammatory cytokine IL-10 with decreased production of proinflammatory cytokines TNF and IL-1β. Phenotypic analysis of M1- and M2-polarized wild-type and mirn23a knockout–derived macrophages demonstrated an inhibition of M1 polarization and an enhancement of M2 macrophages of knockout cells. Finally, to observe an in vivo functional effect on increased M2 polarization of macrophages, we inoculated wild-type and mirn23a−/− mice with two syngeneic tumor cell lines ovarian cancer ID8 and Lewis lung carcinoma (LLC). We observed in vivo that syngeneic tumors grow better in mirn23a−/− mice. The mirn23a−/− increase in tumor growth can be mimicked in wild-type mice by coinjecting mirn23a−/− macrophages with tumor cells, supporting a critical role for mirn23a in macrophage mediate tumor immunity. The data demonstrate a role for mirn23a in macrophage polarization and imply that pharmaceutically targeting mirn23a miRNA in macrophages could be used for programming their polarization/activity. Modulating macrophage inflammatory activity would be useful for improving tumor immunity as well as for the treatment of infectious and autoimmune diseases.

Materials and Methods

Tissue culture

For in vitro macrophage experiment, BMDMs were prepared from bone marrow isolated from wild-type and mirn23a−/− C57BL/6 mice. Briefly, bone marrow was flushed from femurs and tibias of individual mice with sterile PBS (pH 7.4). RBCs were lysed using ammonium–chloride–potassium (ACK) lysis buffer (150 mM NH4Cl, 10 mM KHCO3, and 0.1 mM EDTA in PBS). After ACK lysis, marrow was plated on 10-cm tissue culture plates in DMEM (Sigma-Aldrich, St. Louis, MO), 10% FBS (Peak Serum, Wellington, CO), 50 U/ml penicillin, and 50 μg/ml streptomycin. Unless stated otherwise, cell culture media additives were obtained from Invitrogen (Carlsbad, CA). Cells were incubated at 37°C/5% CO2 for 4 h. After incubation, nonadherent cells were pelleted and resuspended in DMEM, 10% FBS, 15% L929 conditioned media, 10 mM HEPES buffer, and 50 U/ml penicillin, and 50 μg/ml streptomycin and plated for 7 d at 37°C/5% CO2. Fresh media was added on day 4 to supplement old media (no media removed). After 7 d, macrophages were detached using TrypLE (Thermo Fisher Scientific, Pittsburgh, PA), counted, and plated at 1 × 106 cells/ml overnight. The following day, macrophages were stimulated with either 0, 100, or 1000 ng/ml LPS (L8643; derived from Pseudomonas aeruginosa; Sigma-Aldrich); M1-polarization conditions were the following: 100 ng/ml LPS + 50 ng/ml mIFN-γ (M1), or M2-polarizing conditions were the following: 20 ng/ml mIL-4 + 20 ng/ml mIL-13 (M2). Cytokines were obtained from BioLegend (San Diego, CA). RAW264.7 cells were obtained from American Type Culture Collection (TIB-71; Manassas, VA). Cells were maintained in DMEM, 10% FBS, 50 U/ml penicillin, and 50 μg/ml streptomycin. RAW264.7 were polarized or treated with LPS as described for BMDMs. ID8p53−/− (ID8 clone F3) murine ovarian cancer cell lines were provided by I. A. McNeish (University of Glasgow) (30). Cells were cultured in DMEM, 4% FBS, 50 U/ml penicillin, 50 μg/ml streptomycin, and 1× 5 μg/ml insulin, 5 μg/ml transferrin, and 5 ng/ml sodium selenite. LLC cells and EL4 lymphoma cells were obtained from American Type Culture Collection (CRL-1642 and TIB-39) and were cultured in DMEM, 10% FBS, 50 U/ml penicillin, and 50 μg/ml streptomycin to prepare for injection.

Morphology induced by LPS stimulation

To investigate morphological differences, BMDMs were stimulated with LPS (L8643; Sigma-Aldrich) for 6 h. Three mice were analyzed per genotype. After 6 h of stimulation with LPS, four transmitted light fields were taken per mouse at 20× on an EVOS Cell Imaging System (Thermo Fisher Scientific). Cells were considered to have morphology associated with an inflammatory phenotype if they 1) had flattened themselves against the plate and gained an omelet appearance, and 2) if they had a branched morphology designated by lamellipodia and filopodia. Cells that maintained rounded morphology or maintained an elongation phenotype associated with M2 polarization were counted toward the noninflammatory total. The four fields were averaged per mouse as a percent of inflammatory cells to total cells, and then these values were averaged per genotype.

Quantitative RT-PCR

Total RNA was isolated using TRIzol reagent (Thermo Fisher Scientific), according to manufacturer’s protocol. RNA from primary lineage negative cells was isolated using an miRNeasy kit from QIAGEN (Germantown, MD). RNA was reversed transcribed to cDNA using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). Quantitative PCR (qPCR) was performed using gene-specific fluorescent probe assays for both miRNAs and mRNAs on a CFX96 C1000 system (Bio-Rad Laboratories, Hercules, CA). Relative gene expression was calculated using the comparative threshold cycle method. Sno202snRNA and Gapdh were used as housekeeping genes for miRNAs and mRNA, respectively. miRNA cDNA primers (5×) and qPCR primers (20×) were ordered from Thermo Fisher Scientific. All other primers for qPCR were ordered from Integrated DNA Technologies (Coralville, IA). Unless noted otherwise, all q-RTPCR analysis was performed with three or more experimental replicates and three technical replicates.

ELISA for TNF

BMDMs derived from wild-type and mirn23a−/− mice. Cells plated at a density of 1 × 106 cells/ml 1 d prior to use were treated with 1000 ng/ml of LPS. After 1, 6, 12, and 24 h, tissue culture supernatant was isolated for ELISA analysis. The media was subsequently centrifuged to remove any BMDMs that were dislodged, and the supernatant was carefully transferred to a new tube and frozen at −80°C until assayed. A total of 100 μl of conditioned media from (un)stimulated BMDMs was analyzed. BMDM supernatants were analyzed for TNF concentrations with a Ready Set Go ELISA (Thermo Fisher Scientific) kit per manufacture protocol.

Immunoblots

Whole-cell lysates were prepared by lysing cells in radioimmunoprecipitation assay buffer (50 mM Tris [pH 7.5], 150 mM NaCl, 1% NP-40, 0.5% EDTA, and 0.1% SDS) supplemented with 2× Halt protease inhibitor (Pierce, Rockford, IL). Protein was separated by SDS-PAGE on a 10% polyacrylamide gel and transferred to nitrocellulose membrane. Protein concentration was determined using bicinchoninic acid assay (Pierce). Forty to seventy-five micrograms of protein was used per lane. Abs to proteins of interest STAT6 (D3H4), phospho-STAT6 pY641 (D859Y), STAT3 (124H6), and β-actin (8H10D10) (Cell Signaling Technology, Danvers, MA) were used in Western blots. Signals were developed with ECL (Pierce). Analysis was performed on the ChemiDoc XRS+ System using Image Lab (Bio-Rad Laboratories) or ImageJ (U.S. National Institutes of Health, Bethesda, MD) software.

Flow cytometry

The polarization of BMDMs was evaluated by flow cytometry. BMDMs were prepared as described above from the bone of wild-type and mirn23a−/− mice. After 6 d, BMDMs were replated at 500,000 cells per well of a six-well nontissue culture plate. Cells for M0 and M2 polarization were replated in M-CSF (10% L929 conditioned media) containing media, whereas M1-polarized cells were plated in GM-CSF (10% HM5 conditioned media) containing media. Twenty-four hours after replating, M1 cells were treated with IFN-γ and LPS, and M2 cells were treated with IL-4 and IL-13 as described above for 12 or 24 h. Single-cell suspensions were prepared using TrypLE. Cell suspension were incubated with Fc Block (BioLegend) to block nonspecific Ab binding. Cells were then incubated with Abs to extracellular Ags, IA/IE, CD206, and CD38 (BioLegend) in PBS/BSA. For detection of intracellular Ags, cells were subsequently fixed and permeabilized with eBioscience Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific) according to manufacturers’ protocol. After fixation and permeabilization, cells were incubated with fluorochrome conjugated Abs to iNOS, arginase 1, and Egr-2 (Thermo Fisher Scientific). Labeled cells suspensions were analyzed on an FC500 (Beckman Coulter, Pasadena, CA) flow cytometer.

Macrophage killing assay

We plated 500,000 wild-type and mirn23a−/− BMDMs into each well of a six-well plate in 2.5 ml M-CSF (L929) BMDM media. Twenty-four hours later, 1 × 106 EL4 cells per well were added to BMDMs, or as a control, EL4 cells were plated alone. Cells were then cultured in the presence of absence of 20 ng/ml LPS to activate macrophage tumoricidal killing. Forty-eight hours later, the nonadherent EL4 cells were collected and incubated with annexin V–PE for flow cytometric analysis to assay cell viability.

ID8p53−/− coculture with BMDMs

BMDMs were seeded at 1 × 106 cells per well of six-well tissue culture (treated) plates. BMDMs were exposed to an equal number of ID8p53−/− (ID8 F3 Trp53−/− clone) (30) cells across 0.4-μm Transwell inserts (Corning Life Sciences, Corning, NY) for 24 h. RNA was prepared, and quantitative RT-PCR (qRT-PCR) was performed as described previously.

Animals

Mirn23a−/− mice were previously described (18). Wild-type and mirn23a−/− mice between 6 and 8 wk of age were used for all experiments. For generation of BMDMs, mice were euthanized in the Euthanex chamber (Perotech Sciences, Markham, ON, Canada) in the Freimann Life Science Center (University of Notre Dame). Femurs and tibias were then isolated as a source of bone marrow. For tumor survival studies, wild-type and mirn23a−/− C57BL/6 mice were inoculated with ID8p53−/− syngeneic ovarian tumor cells. In brief, 5 × 106 ID8p53−/− cells were injected i.p. into eight wild-type and eight mirn23a−/− mice. For LLC tumor studies, 6–10-wk-old mice were s.c. injected in their lower right flank with 3 × 105 cells in 100 μl PBS. When tumors became palpable (typically 10 d postinjection), tumor growth was monitored every 3 d, and tumors were measured via caliper. Tumor volume was calculated as volume = 0.52 × (width)2 × (length). For BMDM coinjection studies, 2.25 × 105 LLC cells were injected with either 7.5 × 104 wild-type or mirn23a−/− BMDMs (3:1 tumor to BMDM) as described above. For analysis of tumor-associated macrophages (TAMs), resected LLC tumors were mechanically minced and then incubated in RPMI 1640 (Sigma-Aldrich) at 4° with supplemented Collagenase IV (2500 U/ml) (Thermo Fisher Scientific) and DNase I (200 U/ml) (Promega, Madison WI) for 1 h. The resultant single-cell suspensions were strained through 70-μm filters. To obtain the enriched immune cell fraction, the filtered suspension was separated via Ficoll-Paque at 805 × g for 20 min at 20°C. The immune cell fraction was carefully removed and prepared for flow cytometric analysis. Cells were stained with fluorochrome-labeled Abs recognizing CD11b, CD45, CD206, and F4/80 (BioLegend). Labeled cell suspensions were analyzed on an FC500 (Beckman Coulter) flow cytometer.

The University of Notre Dame Institutional Animal Care and Use Committee policy for humane end points in animal experimentation were followed. Mice were visually monitored daily for signs of lethargy and abdominal distension. End points for euthanasia included lethargy, decreased motility, and cross-sectional abdominal diameter increase >33%. The use of mice in these experiments was approved by the Indiana University School of Medicine and University of Notre Dame institutional animal care and use committees (protocols 16-05-3180 [K.D.C.D.] and 16-08-3312 [R.D.]).

RNA-seq

Myeloid progenitors (Lin-kit+sca1−) were sorted from bone marrow on an FACSAria III. ACK-lysed bone marrow was incubated with biotin-labeled Lin mixture Abs (Stem Cell Technologies, Vancouver, CA), streptavidin–APC-Cy7, Sca1-FITC, c-Kit–APC. Sca1 and c-Kit Abs were obtained from BioLegend. For each genotype (wild-type and mirn23a−/−), four separate RNA preparations were prepared and used to build sequencing libraries. Each RNA preparation was prepared from sorted cells acquired from bone marrow isolated from three to four mice. M1/M2-polarized macrophages were prepared as described above. RNA was prepared from BMDMs generated from three unique animals for each genotype. BMDMs were washed two times in PBS (pH 7.4). Cells were subsequently lysed in TRIzol reagent (Thermo Fisher Scientific), and RNA was isolated using Qiagen miRNeasy kit (QIAGEN, Valencia, CA) following manufacturer protocol. RNA-seq libraries were prepared at the University of Notre Dame Genomics and Bioinformatics Core Facility. Total RNA samples were diluted five times prior to analysis. Sample concentration was measured using Qubit RNA HS Assay Kit (Q32855; Invitrogen). Total RNA was evaluated with Agilent Bioanalyzer 2100 System and Agilent RNA 6000 Nano Kit (5067-1511; Agilent Technologies, Santa Clara, CA). Samples with an RNA integrity number of 7 or higher were qualified for library preparation. Total RNA input was normalized to 150 ng. Polyadenylated RNA molecules were selected for using NEBNext Poly(A) mRNA Magnetic Isolation Module (E7490S/L; New England BioLabs, Ipswich, MA). Enriched polyadenylated RNA was converted into an Illumina library using NEBNext Ultra II RNA Library Prep with Sample Purification Beads (E7775S/L; New England BioLabs) and barcoded with NEBNext Multiplex Oligos for Illumina (Index Primers Set 1) (E7335S/L; New England BioLabs) or NEBNext Multiplex Oligos for Illumina (Index Primers Set 2) (E7500S/L; New England BioLabs). Indexed libraries were quantitated with Qubit dsDNA HS Assay Kit (Q32854; Invitrogen). Library quality assessment with Agilent DNA 7500 Kit (5067-1506; Agilent Technologies). The individual libraries were normalized, and equal molar amounts were multiplexed into a single pool. Molar concentration of the multiplex pool was determined with KAPA Library Quantification Kits for Illumina (KK4824; KAPA Biosystems, Boston, MA). Materials used were the following: NEBNext Ultra II RNA Library Prep with Sample Purification Beads (E7775S/L; New England BioLabs), NEBNext Poly(A) mRNA Magnetic Isolation Module (E7490S/L; New England BioLabs), NEBNext Multiplex Oligos for Illumina (Index Primers Set 1) (E7335S/L; New England BioLabs), NEBNext Multiplex Oligos for Illumina (Index Primers Set 2) (E7500S/L; New England BioLabs), Agilent RNA 6000 Nano Kit (5067-1511; Agilent Technologies), Agilent DNA 7500 Kit (5067-1506; Agilent Technologies), Qubit RNA HS Assay Kit (Q32855; Invitrogen), Qubit dsDNA HS Assay Kit (Q32854; Invitrogen), and KAPA Library Quantification Kits for Illumina (KK4824; KAPA Biosystems). Libraries were sequenced on an Illumina NextSeq 500, paired end, 75 total cycles to obtain >20 million reads per sample. Reads were aligned to the human genome (hg19) using STAR (PMC3530905). Gene-level expression was calculated (fragments per kb per million) using Cufflinks (PMC3146043) and differentially expressed genes identified using DESeq (PMC3218662). For all algorithms, default parameters were used unless otherwise noted. Pathway analysis and upstream regulator analysis was performed using Ingenuity Pathway Analysis (IPA; https://www.qiagenbioinformatics.com/products/ingenuitypathway-analysis; QIAGEN), Biojupies (https://biojupies.cloud/), and Broad Institute Gene Set Enrichment Analysis (GSEA) software (http://software.broadinstitute.org/gsea/index.jsp). RNA-seq data are accessible by querying the accession numbers GSE160214 and GSE159762 at National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/).

Statistical analysis

Statistical data are presented as ±SEM. Differences between sample groups were determined by performing an unpaired Student t test. Analysis was performed in Microsoft Excel and using Prism software, version 6.0 (GraphPad Software, La Jolla, CA).

Results

Loss of mirn23a in myeloid progenitor cells downregulates inflammatory signaling

Previously, we observed that mice with homozygous germline null mutations in mirn23a alleles results in skewed differentiation of HSPCs toward B lymphopoiesis at the expense of myelopoiesis (18). Mirn23a miRNAs agonize the PI3K/Akt pathway and antagonize BMP4/Smad signaling, contributing to their promotion of myeloid development (19). To identify additional mirn23a targets that contribute to committing HSPCs to myelopoiesis, we isolated RNA from wild-type and mirn23a−/− myeloid progenitors (Lin−Sca1−Kit+) and performed RNA-seq analysis. Genes that were either significantly upregulated or downregulated 2-fold are shown in Tables I, II, respectively. When we performed pathway analysis on differentially expressed genes, surprisingly, we observed that loss of mirn23a resulted in the downregulation of several inflammatory pathways (Fig. 1). GSEA (31) querying the Hallmark gene sets revealed that inflammatory response genes were enriched in wild-type cells compared with mutant (Fig. 1A). Specifically, IFN classes I and II, TNF, and IL-6 signaling were downregulated in mutant cells as evaluated by GSEA. Consistent with the GSEA, querying the Reactome pathway database (32) demonstrated that TLR signaling and immune system pathways were downregulated in the absence of mirn23a (Fig. 1B). Finally, we analyzed the differentially expressed genes with IPA (QIAGEN) (33), Upstream regulator analysis identifies upstream regulators that putatively explain observed gene expression changes in expression datasets. IPA revealed inhibition of inflammatory signaling regulators in mutant cells (Fig. 1C). Molecules/stimuli/pathways with z-scores <−2.0 are considered to be inhibited. TLR, IFN, and activation of NF-κB transcription complexes were all significantly downregulated in the mirn23a−/− myeloid progenitor datasets. The analysis also revealed downregulation of p38 and JNK stress transducing MAPK activity, which are responsive to inflammatory signals (34).

View this table:
Table I. 2-Fold–upregulated genes between wild-type and mirn23a−/− myeloid progenitors
View this table:
Table II. 2-Fold–downregulated genes between wild-type and mirn23a−/− myeloid progenitors
FIGURE 1.
FIGURE 1.

RNA-seq analysis of RNA isolated from myeloid progenitors reveals that mirn23a loss downregulates inflammatory signaling. RNA was isolated from wild-type (n = 4) and mirn23a−/− (n = 4) bone marrow myeloid progenitors (Lin-Sca1-cKit+) and subjected to RNA-seq analysis. (A) GSEA plots of wild-type versus mirn23a−/− myeloid progenitor gene expression interrogating the Molecular Signatures Database (MSigDB) Hallmark Gene Set. (B) Differentially expressed genes were used to interrogate the Reactome Pathway analysis using BioJupies software. (C) Differentially expressed genes were used to identify upstream regulators affected by loss of mirn23a using IPA software.

Mirn23a−/− decreases inflammatory response to LPS in BMDMs

The RNA-seq data suggested that macrophages deficient in mirn23a have reduced responses to inflammatory signaling. We generated BMDMs from wild-type and mirn23a−/− mice to examine the role of cluster miRNAs in regulating the inflammatory functions of these critical regulators of immune function. In agreement with previous reports (20, 27, 29, 35), we observed that inflammatory molecule/TLR4 ligand, LPS, decreased the expression of all three mirn23a miRNAs in BMDMs after 24 h of treatment (Fig. 2A). When BMDMs from wild type and mirn23a−/− were challenged for 24 h with 0, 100, or 1000 ng/ml of LPS, we observed significantly decreased expression of proinflammatory cytokines Tnf and Il1b with the highest concentration of LPS in mirn23a−/− BMDMs compared with wild-type (Fig. 2B). Consistent with this decrease in production of inflammatory cytokines, we observed, after 6 h LPS stimulation, mirn23a−/− BMDMs have delayed acquisition of an elongated morphology commonly gained upon exposure to inflammatory stimuli compared with wild type (Fig. 2C, 2D) (36, 37).

FIGURE 2.
FIGURE 2.

Mirn23a BMDMs have a reduced response to bacterially derived LPS. (A) qRT-PCR analysis of mirn23a miRNA (miRs-23, -24-2, and -27a) expression in wild-type (Wt) BMDMs treated with LPS for 24 h. (B) Expression of proinflammatory cytokines Tnf and Il1b in LPS-treated Wt (n = 3) and mirn23a−/− (A−/−; n = 3) BMDMs. (C) Representative bright-field images of Wt (n = 3) and A−/− BMDMs (n = 3) treated with 1000 ng/ml LPS for 6 h. Arrows indicate cells displaying an elongated morphology associated with a response to LPS. (D) Quantification of the percentage of cells demonstrating elongated morphology. Error bars represent the SEM. *p < 0.05, **p < 0.01.

Next, we performed a time course of gene expression after LPS treatment of BMDMs. After 6 h of LPS treatment, we observed significantly decreased expression of miRs-23a, -27a, and -24 (Fig. 3A). Downregulation of miR-24 is not as robust likely because of contribution of miR-24 produced from the mirn23a paralog gene mirn23b (the two genes produce identical mature miR-24-3p). In addition to examining miRNA expression, we quantified the expression of proinflammatory cytokines Tnf, Il1b, Il6, Il12b, and anti-inflammatory cytokine Il10 (Fig. 3B). Significant differences for all cytokines were observed after 6 h of treatment. For Tnf and Il1b, there was a significant decrease in expression detected after only 0.5 h. An ELISA was performed on conditioned media for the presence of TNF, and in agreement with gene expression analysis, we observed decreased concentration of TNF in the media obtained from LPS-treated mirn23a−/− cell compared with wild-type BMDMs (Fig. 3C). For the anti-inflammatory gene Il10, we observed more robust expression in mutant macrophages 0.5 h after LPS treatment, which continued for up to 6 h posttreatment (Fig. 3B). Last, protein expression of LPS-induced/anti-inflammatory protein A20 (coded for by Tnfaip3) was examined by immunoblot (Fig. 3D, 3E). A20 is required for terminating TLR signaling in macrophages and protects mice from LPS-induced endotoxic shock (38). Expression of miR-23a inhibitors and mimics have demonstrated A20 to be a miR-23a target in macrophages (39). Basal levels of A20 were ∼2-fold higher in mirn23a−/− BMDMs. Six hours after LPS treatment of wild-type and mirn23a−/− macrophages, we observed induction of A20 in both genotypes. A20 in LPS-induced macrophages was modestly increased (∼1.3-fold) in mutant cells compared with wild type (p < 0.015). The effect of mirn23a loss on cytokine expression and A20 protein levels after LPS treatment of macrophages is consistent with the decreased inflammatory signaling pathways we observed with myeloid progenitors.

FIGURE 3.
FIGURE 3.

Time course of cytokine expression in LPS-treated BMDMs. BMDMs were treated with 1000 ng/ml of LPS for the indicated times and RNA harvested for qRT-PCR analysis. (A) qRT-PCR analysis of mirn23a miRNA relative expression in wild-type (Wt) BMDMs treated with LPS (n = 3). (B) Relative expression of cytokines in LPS-treated Wt (n = 3) and mirn23a−/− (A−/−; n = 3) BMDMs. (C) BMDMs (Wt; n = 3) and A−/− (n = 3) were treated with 1000 ng/ml of LPS for the indicated times and tissue culture supernatant isolated for ELISA analysis. (D) Whole-cell lysates were prepared from Wt (n = 3) and A−/− (n = 3) BMDMs treated with 1 μg/ml LPS for the indicated times. Lysates were subjected to SDS-PAGE for immunoblotting. Blots were probed with Abs to A20 and β-actin. A20 expression normalized to β-actin. (E) Quantitated relative expression of A20 levels in Wt cells prior to LPS stimulation. Error bars represent the SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Mirn23a regulates macrophage polarization

The decreased inflammatory response to LPS suggested that mirn23a miRNAs regulate macrophage polarization. To investigate this, BMDMs were polarized with M1 (LPS and IFN-γ) and M2 (IL-4 and IL-13) conditions for 24 h. To evaluate the polarization status of the cells, expression of Il12b, Nos2, and Arg1 were assayed. IL-12β is a proinflammatory cytokine induced by the M1 stimuli LPS and IFN-γ (40). iNOS (protein product of nos2) and Arg-1 are involved in arginine metabolism (41). When macrophages are polarized to M1, arginine is broken down by iNOS to generate reactive oxygen species, whereas in M2 macrophages, arginine is broken down into ornithine and other metabolites that are used for tissue repair. In accordance with our data from Figs. 2 and 3, mirn23a−/− BMDMs polarized to M1 had significantly decreased expression of proinflammatory genes Il12b and Nos2 compared with wild type (Fig. 4A). Mirn23a−/− M2 polarized BMDMs had significantly increased expression of M2 associated gene Arg1 compared with wild type (Fig. 4B). Mirn23a−/− M1-polarized BMDMs had ∼2-fold decrease in Nos2 compared with wild-type polarized cells, and mutant M2 cells had ∼2-fold increase in Arg1 compared with wild type. We next examined the polarization of wild-type and mirn23a−/− macrophages by flow cytometry assaying expression of iNOS (M1) and Arg-1 (M2) protein (Fig. 4C). BMDMs were prepared from both genotypes and polarized for 12 h. Naive macrophages were not positive for iNOS or Arg-1 expression. No significant genotype specific differences in iNOS-positive cells were observed with M1 polarization, which resulted in over 90% of cells expressing detectable iNOS in both genotypes. However, when cells were M2 polarized, we observed ∼3-fold more Arg-1+ cells in the mirn23a−/− cells compared with wild type (Fig. 4D). Similar results were obtained when we examined polarization after 24 h. We also examined polarization of macrophages with two additional panels CD38 (M1) versus EGR2 (M2) and IA/IE (M1) versus CD206 (M2) (42–44). In agreement with what we observed with iNOS-positive cells, we did not observe any significant difference in the expression of IA/IE or CD38 on M1 BMDMs (Supplemental Fig. 1). However, we did observe ∼2-fold increases in EGR-2+ and CD206+ mirn23a−/− macrophages compared with wild type under M2-polarization conditions. The discrepancy in Nos2 levels being decreased in mirn23a cells as assayed by qRT-PCR, but no difference observed in percent positive iNOS+ cells assayed by flow cytometry, is likely due to decreased abundance in Nos2 transcripts per cell.

FIGURE 4.
FIGURE 4.

Loss of mirn23a reduces M1 polarization and enhances M2 polarization of BMDMs. Wild-type (Wt) and mirn23a−/− (A−/−) BMDMs were polarized to M1 (IFN-γ/LPS) and M2 (IL-4/IL-13) for 24 h. RNA was harvested and analyzed by qRT-PCR for relative expression of (A) M1-associated genes Il12b (n = 3 each genotype) and Nos2 (n = 8 each genotype), and (B) M2-associated gene Arg1 (n = 6 each genotype). (C) Wt (n = 3) and A−/− (n = 3) BMDMs polarized to M1 (24 h pretreat GM-CSF, followed by 12 h IFN-γ/LPS) and M2 (M-CSF/IL-4/IL-13) for 12 h were collected and assayed for expression of M1 protein iNOS and M2 protein Arg-1 by flow cytometry. (D) Average positive cells for iNOS and Arg-1 protein expression. (E) Whole-cell lysates were prepared from Wt and A−/− BMDMs that were unstimulated or stimulated with IL-4 and IL-13 (M2 conditions) for 6 h. Proteins blots were prepared and probed with Abs to total STAT6 and STAT3, phospho-STAT6, and β-actin. (F) Relative total STAT6 and phospho-STAT6 were quantified. Protein loading was normalized to total STAT3, which was not significantly affected by cytokine stimulation. Error bars represent the SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

To further analyze the effect mirn23a loss had on M2 polarization, we examined the activation of STAT6 downstream of the M2-polarizing conditions (IL-4 and IL-13). IL-4 activates STAT6 through receptor activation of JAK3 kinase, which results in phosphorylated STAT6 translocating to the nucleus, where it activates genes associated with mouse M2 macrophages, including Arg1 and macrophage mannose receptor 1 (Mrc1; CD206) (45, 46). STAT6 and JAK3 have been shown to be direct targets of mirn23a miRNAs. Protein extracts were prepared from unstimulated (M0) BMDMs derived from genotypes and BMDMs treated with IL-4 and IL-13 (M2 conditions) for 6 h. Modest, but significant, increases in total STAT6 (p < 0.007) and phospho-STAT6 (p < 0.04) were observed in 6-h, M2-polarized mirn23a−/− BMDMs compared with wild type (Fig. 4E, 4F). There was also a modest increase observed in unstimulated BMDMs, but this was not quite significant, with a p value of <0.08. No changes were observed in total levels of STAT3.

Previously, work investigating the mirn23 cluster only examined macrophage polarization when a single-cluster miRNA is overexpressed or knocked down, and these studies have yielded contradictory results (23–26, 47). In this study, we overexpressed the mirn23a cluster (23aCl) in RAW264.7 murine macrophages. Mirn23a overexpression was confirmed by examining miR-23a, -24-2, and -27a levels. All miRNAs were expressed in the 23aCl cells at least 4-fold above expression in the control cells (stably infected with murine stem cell virus [MSCV]) (Supplemental Fig. 2A). We polarized the MSCV and 23aCL RAW264.7 cells for 24 h with either M1 or M2 conditions. RNA was harvested, and gene expression associated for the M1 and M2 phenotype was examined. M1-polarized 23aCl macrophages had ∼2-fold higher expression of Tnf and Nos2 compared with MSCV cells, whereas M2 polarization resulted in over a 2-fold decrease in Arg1 and an over 80% decrease in Il10 expression (Supplemental Fig. 2B). We also observed that LPS treatment alone of 23aCL macrophages resulted in a decrease in Il10 transcription (Supplemental Fig. 2C).

Mirn23a−/−-polarized BMDMs are enriched for M2 programs

To further characterize mirn23a regulation of macrophage polarization, we performed RNA-seq on M1- and M2-polarized wild-type and mirn23a−/− BMDMs. Transcripts differentially regulated greater or <2-fold (1.0 log2 fold change) in mirn23a−/− M1- and M2-polarized BMDMs are shown in Fig. 5A, 5B. Strikingly, the number of genes with significant (p < 0.05) 2-fold changes in our M1 dataset is 4-fold higher than the 2-fold–changed genes in our M2 dataset. Forty-three unique genes were differentially expressed over 2-fold in M1 mutant macrophages compared with wild type, with 29 upregulated and 14 downregulated. Of the 29 upregulated genes, 10 genes are known to be associated with M2 polarization, with only two associated with M1 polarization. Of the 14 repressed genes, four are reported to be associated with M1 polarization and two (Hmox1 and Ppap2b) with M2 polarization. Although Hmox1 and Ppap2b are described to be associated with M2 polarization, in our hands, they are expressed 8.0- and 4.6-fold higher, respectively, in wild-type M1 cells compared with M2 cells. We found that 10 genes were differentially expressed in mirn23a−/− M2 macrophages by at least 2-fold compared with wild type. Eight of these genes were repressed. Additionally, 4/10 of the M2 differentially expressed were also differentially expressed with M1 conditions. Several of the up- and downregulated targets between the M1 and M2 datasets are the same, suggesting that mirn23a regulates general macrophage functions as well as polarization specific functions.

FIGURE 5.
FIGURE 5.

Pathway analysis of RNA-seq data obtained from wild-type and mirn23a−/− M2-polarized BMDMs. RNA-seq analysis was performed on RNA isolated from wild-type (n = 3) and mirn23a−/− BMDMs (n = 3) after polarization for 24 h with M2 conditions. 2-Fold differentially expressed genes between wild-type and mirn23a−/− BMDMs. (A) M1 polarized and (B) M2 polarized. Red/green indicate up- and downregulated genes, respectively. Dark and light shades indicate gene product association with M2 or M1 polarizations, respectively. Intermediate shading indicates that nothing in the literature supports a role in macrophage polarization. (C) IPA canonical pathway analysis. Pathways with a p < 0.05 and with a z-score ≥1.0 or ≤−1.0 are shown. Red bars indicate predicted upregulated pathways, and green bars indicate downregulated pathways in mirn23a−/− M2 macrophages. (D) Top 16 downregulated Reactome pathways in mirn23a−/− M2 macrophages compared with wild-type macrophages ranked by p value.

We performed canonical pathway analysis with IPA software (QIAGEN) to analyze differentially expressed genes in M1-polarized macrophages. Pathways with a p < 0.05 and with a z-score greater than or equal to 1.0 or less than or equal to −1.0 are shown in Fig. 5C. As evaluated by z-score, the most highly activated pathway is lymphotoxin β receptor signaling. Activation of lymphotoxin β receptor downregulates inflammatory signaling through inducing cross-tolerance to TLR4 and TLR9 ligands (48). The two most downregulated pathways were ceramide and p53 signaling. Ceramide is composed of sphingosine and a fatty acid that can act as a cell signaling molecule through activation of phosphatase 2A, p38, JNK, AKT, protein kinase Cζ, and survivin. It was shown to have anti-inflammatory and proinflammatory properties. p53 is known for its role as a tumor suppressor; however, it also regulates inflammatory signaling in macrophages (49). P53 activity increases when macrophages are polarized to the M2, leading to reduced expression of M2 genes. Macrophages lacking p53 express higher levels of M2 genes during M2 polarization (49).

Differential gene expression between wild-type and mirn23a−/− M1-polarized macrophages was examined for enrichment or de-enrichment of Reactome pathways. Of the top 16 significantly downregulated pathways in mutant M1 macrophages compared with wild type, we observed that seven of these pathways were related to TLR signaling (Fig. 5D). The downregulation of TLR pathways is similar to what we observed in myeloid progenitors freshly isolated from bone marrow (Fig. 1B).

Physiological consequences of mirn23a expression

In vitro evaluation of macrophages revealed that mirn23a miRNAs repress or limit M2 polarization. It is not clear though if these effects on polarization are significant enough to affect macrophage function in vivo. Because of their inflammatory plasticity, macrophages are involved in many disease states, including antitumor immunity (5, 50–52). In the tumor microenvironment, TAMs acquire a M2-like phenotype that promotes tumor growth and migration, as well as suppressing the local immune environment to assist the tumor in evading an immune response (51). Increased M2-polarized TAMs in the tumor microenvironment is associated with poorer prognosis for several tumor types (53). Tumor cells secrete factors that polarize macrophages to an immunosuppressive M2 phenotype (54). M2 polarization of TAMs contributes to the progression of tumor growth in the ID8 syngeneic murine tumor model of ovarian cancer (55). Reprogramming of macrophages to an M1 phenotype restricts the growth of ID8 tumors and ablating macrophages in the ID8 tumor microenvironment slows tumor growth. In our study, we used an ID8 line deficient in Trp53 (ID8p53−/−) that elicits an immune response similar to human high grade serous ovarian cancer with increased CD11b+ myeloid cell infiltrates compared with parental ID8 cells (30). Several different tumor cell lines including ID8 have been shown to polarize macrophages to an anti-inflammatory state through the release of cytokines and exosomes (54, 56–59). To determine if loss of mirn23a expression in macrophages enhances M2 polarization by tumor cells as we have observed with recombinant cytokines, we performed Transwell assays with ID8p53−/− mouse ovarian cancer cells and BMDMs. Cells in the Transwell assay do not have direct contact and therefore are influenced by secreted factors. BMDMs from wild-type and mirn23a−/− mice were incubated with ID8p53−/− cells for 24 h. Cocultured macrophages had no significant differences in expression of Tnf or Tgfb; however, there was significantly increased expression of the potent anti-inflammatory cytokine Il-10 (Fig. 6A). These data suggest that regardless of the stimuli, mirn23a−/− macrophages have a bias toward anti-inflammatory M2-polarization states.

FIGURE 6.
FIGURE 6.

Syngeneic tumors generated in mirn23a−/− (A−/−) mice are more aggressive than tumors generated in wild-type (Wt) mice. (A) Wt (n = 3) and A−/− (n = 3) BMDMs were plated onto six-well plates. Transwell inserts with ID8p53−/− mouse syngeneic ovarian cancers were placed in the wells so soluble factors could diffuse. After 24 h of culture RNA was prepared from BMDMs and qRT-PCR used to analyze expression of the indicated cytokines. (B) Wt (n = 6) and A−/− (n = 6) were subjected to i.p. injection of ID8p53−/− tumor cells. Survival of mice was observed, and a Kaplan–Meier curve was generated. (C) LLC were s.c. injected into the flanks of Wt (n = 6) and A−/− (n = 8) mice. Tumor volume was measure by calipers after injection. (D) After 21 d, tumors were isolated and TAM populations assayed by flow cytometry using cell surface markers CD45, CD11b, F4/80, and CD206. (E) In a second independent experiment, Wt (n = 7) and A−/− (n = 7) mice were s.c. injected with LLC cells and tumors isolated and weighed after 21 d. Average weight of tumors isolated from Wt and mutant mice is shown. (F) Images of LLC tumors isolated from individual mice. Significance of survival data determined by log-rank test. Other data examined by Student t test. Error bars represent the SEM. *p < 0.05, **p < 0.01.

To determine if the increased M2 polarization of mirn23a−/− macrophages by ID8p53−/−cells impacts tumor growth, we generated syngeneic tumors using ID8p53−/− cells in wild-type and mirn23a−/− mice with a C57BL/6 background. Because we demonstrated that mirn23a expression impacts macrophage polarization and other immune cells, we anticipated that tumor growth or burden would be impacted by mirn23a expression. Specifically, we hypothesized that tumor growth in mirn23a−/− mice would decrease live span compared with their wild-type counterparts because of immune suppression. Six mice for each genotype were examined. Mice were euthanized when tumor burden was high, and mice were moribund. Mirn23a−/− mice exhibited a median survival of 51 d; the mean survival for wild-type mice was 57.5 d (Fig. 6B). A log-rank (Mantel–Cox) test determined that this difference was significant, with a p value of 0.0008. Additionally, a Gehan–Breslow–Wilcoxon test determined that the difference was significant with a p value of 0.0015. These results indicate that the presence of mirn23a in the tumor microenvironment impacts ovarian tumor growth.

The i.p. injected ID8p53−/− mimics ovarian cancer metastasis as the tumor cells spread through the peritoneal cavity and hundreds of tumors develop on the abdominal membranes and organs. Additionally, mice with ID8p53 tumors develop extensive ascites; therefore, quantitating tumor mass or volume is difficult, so we performed a s.c. tumor model to better examine tumor growth. LLC cells were injected s.c. into wild-type and mirn23a−/− mice, and the tumor growth was monitored. LLC tumors grew quicker in mirn23a−/− mice with significantly increased tumor volumes compared with wild type at days 19 and 21 (Fig. 6C). Mice were euthanized at day 21 and tumors collected to evaluate size and prepare single-cell suspensions for flow cytometric analysis. Single-cell tumor suspensions demonstrated an increased percentage of CD11b+F4/80+CD206+ macrophages in the immune cell fraction of tumors generated in mirn23a−/− mice (Fig. 4C). CD206 is a marker of M2 macrophages. Tumors isolated from mirn23a−/− mice had an average weight ∼2-fold higher compared with tumors isolated from wild-type animals (Fig. 6E, 6F).

mirn23a−/− macrophages promote tumor growth

Increased TAMs in mirn23a-derived tumors suggested that the more aggressive tumor growth could be due (or partially due) to loss of mirn23a in macrophages. To examine this, we first evaluate the tumor killing activity of mirn23a−/− macrophages. Wild-type and mutant BMDMs cocultured with syngeneic EL4 lymphoma cells in the presence and absence of LPS, which activates the cytotoxic activity of the macrophages (60). After 72 h, the EL4 suspension cells were removed and cell death assayed by annexin V binding using flow cytometry. In the absence of LPS, cell death was similar between EL4s cultured in the absence of BMDMs and cells cocultured with either wild-type or mirn23a−/− BMDMs (Fig. 7A). LPS had no effect on EL4 cell cultured alone. However, LPS-induced ∼8-fold increase in EL4 cell death mediated by wild-type BMDMs but only ∼2-fold increase in EL4 cell death in mirn23a−/− BMDM cocultures. The results suggest that mirn23a−/− macrophages have reduced tumor cell killing activity compared with wild type.

FIGURE 7.
FIGURE 7.

Mirn23a macrophages enhance tumor growth. (A) Nonadherent EL4 cells were plated alone (three technical replicates) or with wild-type (Wt; n = 3) and mirn23a−/− (A−/−; n = 3) BMDMs in the presence and absence of LPS. After 48 h of coculture, EL4 cells were removed and assayed for cell death by annexin V binding. (B) Wt mice were s.c. injected with LLC cells and either Wt (n = 6) or A−/− (n = 6) BMDMs. Tumor volume was measured using calipers over time. (C) Twenty-one days after injection, tumors were isolated and weighed. Average weight of tumors generated in the presence of Wt and A−/− BMDMs is shown. *p < 0.05, ***p < 0.001, #p = 0.053.

We next examined whether mirn23a−/− macrophages alone could enhance the growth of syngeneic tumors. Several studies have previously shown that coinjection of BMDMs with syngeneic tumor cells affect tumor growth in vivo (61–63). M2 macrophages specifically have been shown to promote the in vivo growth of syngeneic LLC tumors (64). LLC cells were coinjected s.c. with either wild-type or mirn23a−/− BMDMs into wild-type C57BL/6 mice. Tumor cell/macrophage-injected mice were examined as previously done for LLC injections into wild-type and mutant mice. Tumor volumes were observed to be significantly increased in mice receiving mirn23a−/− macrophages at days 14, 18, and 21. At 21 d postinjection, tumors were isolated and weighed. Tumors derived in mice receiving mirn23a−/− BMDMs on average weighed more than tumors derived from mice receiving wild-type BMDMs. The difference in average weight was not quite significantly different, with a p value of 0.064. These results imply that mirn23a deficiency in macrophages contributes to the enhanced tumor growth we observed when we transplanted LLC cells into mirn23a−/− mice (Fig. 6).

Discussion

RNA-seq data demonstrated that myeloid progenitor cells lacking the mirn23a miRNA cluster have decreased inflammatory signaling (TLR, type I and II IFNs, TNF, and IL-6). Because these pathways are critical for the function of macrophages, we investigated mirn23a regulation of macrophage inflammatory responses and polarization. Previous work investigating the role of individual mirn23a miRNAs (miRs-23a, -24-2, and -27a) in regulating the inflammatory/polarization state of macrophages used overexpression and underexpression techniques, yielding contradictory results (20–24). To clarify the role of mirn23a miRNAs in regulating macrophage function, we examined BMDMs prepared from wild-type and mirn23a knockout mice. Consistent with the RNA-seq analysis of myeloid progenitor, we observed that LPS treatment of wild-type and mirn23a−/− BMDMs resulted in decreased upregulation of inflammatory cytokines Il1b, Tnf, Il6, and Il12b. In addition, we observed upregulation of the anti-inflammatory cytokine Il10. Similarly, when we polarized macrophages under M1 (IFN-γ and LPS) and M2 (IL-4 and IL-13) conditions, we observed decreased expression of M1-associated genes and increased expression of M2 genes, respectively, by mutant cells. When we overexpress the mirn23a cluster in RAW264.7 macrophages, cluster overexpression increased expression of M1-associated genes (Tnf and Nos2) and decreased expression of M2 genes (Arg1 and Il10) when polarized under M1 and M2 conditions, respectively. The data suggest that mirn23a miRNAs act as repressors of M2 polarization and augment inflammatory signaling required for M1 polarization.

Interestingly, endogenous expression of mirn23a miRNAs is repressed with LPS treatment (Figs. 2A, 3A). Similarly, shifting BMDM from M-CSF media, which primes M2 polarization to GM-CSF–containing media, which primes M1 polarization, results in a ∼2-fold decrease in miRs-23a, -24, and -27a (Supplemental Fig. 1E). In addition, mirn23a miRNAs are increased with stimulation by the M2-polarizing cytokine IL-4 (27). Unlike polarized Th cells, macrophages are not committed to a specific polarization (65). They remain plastic and can be repolarized. This is biologically seen with TAMs, which can enter a tumor as M1-polarized macrophages but respond to tumor-secreted factors by repolarizing to an M2 phenotype. Our data suggest that the mirn23a miRNAs are involved in maintaining the plasticity of polarized macrophages. We propose that in M2-polarized macrophages, mirn23a miRNAs are induced to buffer the expression of M2 genes to maintain protein levels within a specific window of expression. Preventing the cell from obtaining high levels of these M2 factors contributes to the ability of the M2 cells to be repolarized to an M1 phenotype. In the absence of mirn23a, M2 cells are more prone to being locked in to the M2 phenotype. Conversely, in M1-polarized cells, mirn23a is repressed, which makes it easier for M1 cells to re-express M2-associated genes and allow for repolarization to M2. This may explain why we observed greater changes in gene expression in mutant M1 macrophages compared with M2 in the RNA-seq analysis. We derived our macrophages culturing bone marrow in the presence of M-CSF, which can skew polarization of macrophages to M2 in part because of increased activation of FoxO1 (66). We propose that normally mirn23a miRNAs are induced so that M2 cells are able to repolarize to M1. In the absence of mirn23a, it is more difficult to repolarize mirn23a−/− M2-skewed cells to M1, and thus, we observe an increased in the number of differentially expressed cells between wild-type and mutant macrophages treated with IFN-γ and LPS.

LPS-induced downregulation of mirn23a is mediated by transcriptional repression induced by NF-κB protein p65 (35). LPS signaling through TLR4 is a well-recognized mechanism of NF-κB activation leading to expression of inflammatory genes. Unrestrained inflammatory activity can lead to tissue damage and autoimmunity, so to downregulate NF-κB induces a negative feedback loop through transcribing its inhibitors Nfkbia (Iκbα) and Tnfaip3 (A20). Mice lacking Tnfaip3 die of an hyperinflammatory response (67). MiR-23a directly target Tnfaip3 transcripts to reduce expression of A20 protein, which supports the conclusion that LPS/p65 repression of mirn23a is part of the NF-κB feedback mechanism to limit the inflammatory response (39, 47). Our data support that mirn23a physiologically regulates NF-κB signaling, as we observed decreases in Il-1b, Il-12b, Tnf, and Nos2, well-known genes regulated by NF-κB. A20 protein levels are higher in mirn23a−/− BMDMs, which likely contributes to the downregulation of these inflammatory genes in mutant macrophages (Fig. 3D).

The effects on in vitro macrophage responses to LPS and polarization were significant but fairly modest. However, in vitro stimulation of cells is likely stronger than the signals observed in vivo, so it was important to examine if loss of mirn23a had repercussions in vivo in which we could observe cell responses to physiological stimuli. To demonstrate the in vivo significance of the mirn23a-mediated macrophage polarization effects we have observed in vitro, we generated syngeneic tumors in wild-type and mutant mice. M2 polarization of macrophages by tumor cells contributes to the growth of ovarian tumor ID8 cells and LLC cells in vivo (54, 55, 64). Inhibiting macrophage M2 polarization results in decreased growth of ID8 and LLC tumors in vivo. We hypothesized that because the absence of mirn23a skewed macrophages toward M2 polarization in vitro that tumors would grow faster in mutant mice compared with wild type. Consistent with our hypothesis, mirn23a mice succumbed to i.p. ID8 tumors sooner than wild-type mice, and they had increased LLC tumor growth when monitored over time. LLC tumors generated in mutant mice had increased infiltration of F4/80+CD206+ macrophages. Although macrophages are critical in mediating the immune response to ID8- and LLC-derived tumors, it is possible that the decreased survival was due to loss of expression in other immune cell types. Mirn23a-deficient Th cells skew toward Th2 polarization, and mirn23a miRNAs regulate T regulatory cell function. Additionally, our group has published that mirn23a−/− NK cells have reduced production of inflammatory cytokines and have reduced antitumor activity (68). We examined mirn23a−/− macrophage contribution to increased LLC tumor growth in mutant animals by coinjecting wild-type mice with LLC cells and either wild-type or mirn23a−/− BMDMs. This experiment demonstrated that mirn23a deficiency in macrophages alone could potentiate tumor growth, although it does not rule out contribution in other immune cells.

Leveraging miRNAs as a therapeutic option in inflammatory disease is attractive in part because miRNAs fine-tune gene expression. Targeting miRNAs may be an easier way to modulate inflammation as opposed to other pharmaceutical methods that may ablate the inflammatory response. Having this control could be useful in complex inflammatory diseases such as cancer. Tumors manipulate local immune cells such as macrophages to suppress immune responses and evade the immune system (69). These same cells maintain a low level of inflammatory cytokines that continue to promote tumor development. However, when these immune cells maintain a more inflammatory phenotype, the immune response combats the tumor. However, too much inflammation can produce systemic responses that may be harmful to the host. Therefore, precise localized regulation of inflammation in the tumor microenvironment has clinical benefits. Our data suggest that mirn23a expression in the tumor microenvironment regulates survival in response to tumors, setting a proof of principle that mirn23a can be manipulated to influence disease progression. Recently, it was demonstrated that delivery of miR-182 to glioblastoma cells enhances tumor susceptibility to chemotherapy (70). Additionally, lentiviral coexpression of multiple miRNAs (MiRs-124, -128, and -137) in a glioblastoma cell line synergized to increase survival in transplanted mice as well as increase susceptibility to chemotherapy in mouse models (71). Potentially using mixtures of miRNAs mimics and antagonist targeted to macrophages will allow for specific polarization and/or repolarization of macrophages. Several studies demonstrate successful delivery of miRNAs to macrophages in vivo using functionalized nanoparticles in mouse models (72–74). By deciphering the extent to which different miRNAs regulate inflammatory pathways and which pathways they specifically regulate, miRNA combinatorial therapies can be designed for a variety of inflammatory diseases and be used for personalized therapy.

Disclosures

The authors have no financial conflicts of interest.

Acknowledgments

We thank the staff of the University of Notre Dame Freimann Life Sciences Center for mouse breeding and husbandry. We thank Melissa Stephens, Joseph Soros, and the staff of the University of Notre Dame Genomics and Bioinformatic Core Facility for preparation of sequencing libraries and data analysis. Finally, we thank former colleagues Drs. Suzanne S. Bohlson (University of California, Irvine, CA) and Manuel D. Galvan (National Jewish Health) for helpful discussions at the beginning of this project.

Footnotes

  • This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases DK109051 (to R.D.), the Ralph W. and Grace M. Showalter Research Trust Fund (to R.D.), the Indiana Clinical and Translational Sciences Institute (to R.D.), National Cancer Institute CA204231 (to S.R.), the Walther Cancer Foundation Interdisciplinary Interface Training Program (to A.B.), and the Research Like a Champion Program, University of Notre Dame (to A.B.). This work was also supported by the Indiana University School of Medicine South Bend Imaging and Flow Cytometry Core Facility.

  • The sequences presented in this article have been submitted to the National Center for Biotechnology Information Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/) under accession numbers GSE160214 and GSE159762.

  • The online version of this article contains supplemental material.

  • Abbreviations used in this article:

    ACK
    ammonium–chloride–potassium
    BMDM
    bone marrow–derived macrophage
    GSEA
    Gene Set Enrichment Analysis
    HSPC
    hematopoietic stem and progenitor cell
    iNOS
    inducible NO synthase
    IPA
    Ingenuity Pathway Analysis
    LLC
    Lewis lung carcinoma
    miRNA
    microRNA
    MSCV
    murine stem cell virus
    qPCR
    quantitative PCR
    qRT-PCR
    quantitative RT-PCR
    RNA-seq
    RNA sequencing
    TAM
    tumor-associated macrophage.

  • Received October 23, 2019.
  • Accepted November 17, 2020.

References

  1. ↵
    1. Franken, L.,
    2. M. Schiwon,
    3. C. Kurts
    . 2016. Macrophages: sentinels and regulators of the immune system. Cell. Microbiol. 18: 475–487.
    OpenUrlCrossRef
  2. ↵
    1. Theret, M.,
    2. R. Mounier,
    3. F. Rossi
    . 2019. The origins and non-canonical functions of macrophages in development and regeneration. Development 146: dev156000.
  3. ↵
    1. Watanabe, S.,
    2. M. Alexander,
    3. A. V. Misharin,
    4. G. R. S. Budinger
    . 2019. The role of macrophages in the resolution of inflammation. J. Clin. Invest. 129: 2619–2628.
    OpenUrlCrossRefPubMed
  4. ↵
    1. Ponzoni, M.,
    2. F. Pastorino,
    3. D. Di Paolo,
    4. P. Perri,
    5. C. Brignole
    . 2018. Targeting macrophages as a potential therapeutic intervention: impact on inflammatory diseases and cancer. Int. J. Mol. Sci. 19: 1953.
    OpenUrl
  5. ↵
    1. Atri, C.,
    2. F. Z. Guerfali,
    3. D. Laouini
    . 2018. Role of human macrophage polarization in inflammation during infectious diseases. Int. J. Mol. Sci. 19: 1801.
    OpenUrl
  6. ↵
    1. Liu, Y. C.,
    2. X. B. Zou,
    3. Y. F. Chai,
    4. Y. M. Yao
    . 2014. Macrophage polarization in inflammatory diseases. Int. J. Biol. Sci. 10: 520–529.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Platanitis, E.,
    2. T. Decker
    . 2018. Regulatory networks involving STATs, IRFs, and NFκB in inflammation. Front. Immunol. 9: 2542.
    OpenUrl
  8. ↵
    1. de Groot, A. E.,
    2. K. J. Pienta
    . 2018. Epigenetic control of macrophage polarization: implications for targeting tumor-associated macrophages. Oncotarget 9: 20908–20927.
    OpenUrl
  9. ↵
    1. Kruidenier, L.,
    2. C. W. Chung,
    3. Z. Cheng,
    4. J. Liddle,
    5. K. Che,
    6. G. Joberty,
    7. M. Bantscheff,
    8. C. Bountra,
    9. A. Bridges,
    10. H. Diallo, et al
    . 2012. A selective jumonji H3K27 demethylase inhibitor modulates the proinflammatory macrophage response. Nature 488: 404–408.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Essandoh, K.,
    2. Y. Li,
    3. J. Huo,
    4. G. C. Fan
    . 2016. MiRNA-mediated macrophage polarization and its potential role in the regulation of inflammatory response. Shock 46: 122–131.
    OpenUrlCrossRefPubMed
  11. ↵
    1. Bartel, D. P.
    2004. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116: 281–297.
    OpenUrlCrossRefPubMed
  12. ↵
    1. Zhang, Y.,
    2. M. Zhang,
    3. M. Zhong,
    4. Q. Suo,
    5. K. Lv
    . 2013. Expression profiles of miRNAs in polarized macrophages. Int. J. Mol. Med. 31: 797–802.
    OpenUrlCrossRefPubMed
  13. ↵
    1. Li, J.,
    2. J. Zhang,
    3. H. Guo,
    4. S. Yang,
    5. W. Fan,
    6. N. Ye,
    7. Z. Tian,
    8. T. Yu,
    9. G. Ai,
    10. Z. Shen, et al
    . 2018. Critical role of alternative M2 skewing in miR-155 deletion-mediated protection of colitis. [Published erratum appears in 2020 Front. Immunol. 10: 3153.] Front. Immunol. 9: 904.
    OpenUrl
  14. ↵
    1. Zhang, Y.,
    2. H. Mei,
    3. X. Chang,
    4. F. Chen,
    5. Y. Zhu,
    6. X. Han
    . 2016. Adipocyte-derived microvesicles from obese mice induce M1 macrophage phenotype through secreted miR-155. J. Mol. Cell Biol. 8: 505–517.
    OpenUrlCrossRefPubMed
  15. ↵
    1. Boldin, M. P.,
    2. K. D. Taganov,
    3. D. S. Rao,
    4. L. Yang,
    5. J. L. Zhao,
    6. M. Kalwani,
    7. Y. Garcia-Flores,
    8. M. Luong,
    9. A. Devrekanli,
    10. J. Xu, et al
    . 2011. miR-146a is a significant brake on autoimmunity, myeloproliferation, and cancer in mice. J. Exp. Med. 208: 1189–1201.
    OpenUrlAbstract/FREE Full Text
  16. ↵
    1. Li, D.,
    2. M. Duan,
    3. Y. Feng,
    4. L. Geng,
    5. X. Li,
    6. W. Zhang
    . 2016. MiR-146a modulates macrophage polarization in systemic juvenile idiopathic arthritis by targeting INHBA. [Published erratum appears in 2017 Mol. Immunol. 87: 329–330.] Mol. Immunol. 77: 205–212.
    OpenUrl
  17. ↵
    1. Kong, K. Y.,
    2. K. S. Owens,
    3. J. H. Rogers,
    4. J. Mullenix,
    5. C. S. Velu,
    6. H. L. Grimes,
    7. R. Dahl
    . 2010. MIR-23A microRNA cluster inhibits B-cell development. Exp. Hematol. 38: 629–640.e1.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Kurkewich, J. L.,
    2. E. Bikorimana,
    3. T. Nguyen,
    4. N. Klopfenstein,
    5. H. Zhang,
    6. W. M. Hallas,
    7. G. Stayback,
    8. M. A. McDowell,
    9. R. Dahl
    . 2016. The mirn23a microRNA cluster antagonizes B cell development. J. Leukoc. Biol. 100: 665–677.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Kurkewich, J. L.,
    2. J. Hansen,
    3. N. Klopfenstein,
    4. H. Zhang,
    5. C. Wood,
    6. A. Boucher,
    7. J. Hickman,
    8. D. E. Muench,
    9. H. L. Grimes,
    10. R. Dahl
    . 2017. The miR-23a∼27a∼24-2 microRNA cluster buffers transcription and signaling pathways during hematopoiesis. PLoS Genet. 13: e1006887.
  20. ↵
    1. Xie, N.,
    2. H. Cui,
    3. S. Banerjee,
    4. Z. Tan,
    5. R. Salomao,
    6. M. Fu,
    7. E. Abraham,
    8. V. J. Thannickal,
    9. G. Liu
    . 2014. miR-27a regulates inflammatory response of macrophages by targeting IL-10. J. Immunol. 193: 327–334.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Hussain, T.,
    2. D. Zhao,
    3. S. Z. A. Shah,
    4. J. Wang,
    5. R. Yue,
    6. Y. Liao,
    7. N. Sabir,
    8. L. Yang,
    9. X. Zhou
    . 2018. MicroRNA 27a-3p regulates antimicrobial responses of murine macrophages infected by Mycobacterium avium subspecies paratuberculosis by targeting interleukin-10 and TGF-β-activated protein kinase 1 binding protein 2. Front. Immunol. 8: 1915.
    OpenUrlCrossRef
  22. ↵
    1. Zheng, Q.,
    2. J. Hou,
    3. Y. Zhou,
    4. Y. Yang,
    5. X. Cao
    . 2016. Type I IFN-inducible downregulation of microRNA-27a feedback inhibits antiviral innate response by upregulating siglec1/TRIM27. J. Immunol. 196: 1317–1326.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Saha, B.,
    2. F. Momen-Heravi,
    3. K. Kodys,
    4. G. Szabo
    . 2016. MicroRNA cargo of extracellular vesicles from alcohol-exposed monocytes signals naive monocytes to differentiate into M2 macrophages. J. Biol. Chem. 291: 149–159.
    OpenUrlAbstract/FREE Full Text
  24. ↵
    1. Saha, B.,
    2. J. C. Bruneau,
    3. K. Kodys,
    4. G. Szabo
    . 2015. Alcohol-induced miR-27a regulates differentiation and M2 macrophage polarization of normal human monocytes. J. Immunol. 194: 3079–3087.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    1. Jingjing, Z.,
    2. Z. Nan,
    3. W. Wei,
    4. G. Qinghe,
    5. W. Weijuan,
    6. W. Peng,
    7. W. Xiangpeng
    . 2017. MicroRNA-24 modulates Staphylococcus aureus-induced macrophage polarization by suppressing CHI3L1. Inflammation 40: 995–1005.
    OpenUrl
  26. ↵
    1. Naqvi, A. R.,
    2. J. B. Fordham,
    3. B. Ganesh,
    4. S. Nares
    . 2016. miR-24, miR-30b and miR-142-3p interfere with antigen processing and presentation by primary macrophages and dendritic cells. Sci. Rep. 6: 32925.
    OpenUrl
  27. ↵
    1. Ma, S.,
    2. M. Liu,
    3. Z. Xu,
    4. Y. Li,
    5. H. Guo,
    6. Y. Ge,
    7. Y. Liu,
    8. D. Zheng,
    9. J. Shi
    . 2016. A double feedback loop mediated by microRNA-23a/27a/24-2 regulates M1 versus M2 macrophage polarization and thus regulates cancer progression. Oncotarget 7: 13502–13519.
    OpenUrlPubMed
  28. ↵
    1. Ma, Y.,
    2. N. Yao,
    3. G. Liu,
    4. L. Dong,
    5. Y. Liu,
    6. M. Zhang,
    7. F. Wang,
    8. B. Wang,
    9. X. Wei,
    10. H. Dong, et al
    . 2015. Functional screen reveals essential roles of miR-27a/24 in differentiation of embryonic stem cells. EMBO J. 34: 361–378.
    OpenUrlAbstract/FREE Full Text
  29. ↵
    1. Si, X.,
    2. D. Cao,
    3. J. Chen,
    4. Y. Nie,
    5. Z. Jiang,
    6. M. Y. Chen,
    7. J. F. Wu,
    8. X. D. Guan
    . 2018. miR-23a downregulation modulates the inflammatory response by targeting ATG12-mediated autophagy. Mol. Med. Rep. 18: 1524–1530.
    OpenUrl
  30. ↵
    1. Walton, J.,
    2. J. Blagih,
    3. D. Ennis,
    4. E. Leung,
    5. S. Dowson,
    6. M. Farquharson,
    7. L. A. Tookman,
    8. C. Orange,
    9. D. Athineos,
    10. S. Mason, et al
    . 2016. CRISPR/Cas9-Mediated Trp53 and Brca2 knockout to generate improved murine models of ovarian high-grade serous carcinoma. Cancer Res. 76: 6118–6129.
    OpenUrlAbstract/FREE Full Text
  31. ↵
    1. Subramanian, A.,
    2. P. Tamayo,
    3. V. K. Mootha,
    4. S. Mukherjee,
    5. B. L. Ebert,
    6. M. A. Gillette,
    7. A. Paulovich,
    8. S. L. Pomeroy,
    9. T. R. Golub,
    10. E. S. Lander,
    11. J. P. Mesirov
    . 2005. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA 102: 15545–15550.
    OpenUrlAbstract/FREE Full Text
  32. ↵
    1. Fabregat, A.,
    2. S. Jupe,
    3. L. Matthews,
    4. K. Sidiropoulos,
    5. M. Gillespie,
    6. P. Garapati,
    7. R. Haw,
    8. B. Jassal,
    9. F. Korninger,
    10. B. May, et al
    . 2018. The reactome pathway knowledgebase. Nucleic Acids Res. 46(D1): D649–D655.
    OpenUrlCrossRefPubMed
  33. ↵
    1. Krämer, A.,
    2. J. Green,
    3. J. Pollard Jr..,
    4. S. Tugendreich
    . 2014. Causal analysis approaches in ingenuity pathway analysis. Bioinformatics 30: 523–530.
    OpenUrlCrossRefPubMed
  34. ↵
    1. Arthur, J. S.,
    2. S. C. Ley
    . 2013. Mitogen-activated protein kinases in innate immunity. Nat. Rev. Immunol. 13: 679–692.
    OpenUrlCrossRefPubMed
  35. ↵
    1. Rathore, M. G.,
    2. A. Saumet,
    3. J. F. Rossi,
    4. C. de Bettignies,
    5. D. Tempé,
    6. C. H. Lecellier,
    7. M. Villalba
    . 2012. The NF-κB member p65 controls glutamine metabolism through miR-23a. Int. J. Biochem. Cell Biol. 44: 1448–1456.
    OpenUrlCrossRefPubMed
  36. ↵
    1. Amano, F.,
    2. Y. Akamatsu
    . 1991. A lipopolysaccharide (LPS)-resistant mutant isolated from a macrophagelike cell line, J774.1, exhibits an altered activated-macrophage phenotype in response to LPS. Infect. Immun. 59: 2166–2174.
    OpenUrlAbstract/FREE Full Text
  37. ↵
    1. Williams, L. M.,
    2. A. J. Ridley
    . 2000. Lipopolysaccharide induces actin reorganization and tyrosine phosphorylation of Pyk2 and paxillin in monocytes and macrophages. J. Immunol. 164: 2028–2036.
    OpenUrlAbstract/FREE Full Text
  38. ↵
    1. Boone, D. L.,
    2. E. E. Turer,
    3. E. G. Lee,
    4. R. C. Ahmad,
    5. M. T. Wheeler,
    6. C. Tsui,
    7. P. Hurley,
    8. M. Chien,
    9. S. Chai,
    10. O. Hitotsumatsu, et al
    . 2004. The ubiquitin-modifying enzyme A20 is required for termination of toll-like receptor responses. [Published erratum appears in 2005 Nat. Immunol. 6: 114.] Nat. Immunol. 5: 1052–1060.
    OpenUrlCrossRefPubMed
  39. ↵
    1. Peng, P.,
    2. Z. Li,
    3. X. Liu
    . 2015. Reduced expression of miR-23a suppresses A20 in TLR-stimulated macrophages. Inflammation 38: 1787–1793.
    OpenUrl
  40. ↵
    1. Ma, X.,
    2. W. Yan,
    3. H. Zheng,
    4. Q. Du,
    5. L. Zhang,
    6. Y. Ban,
    7. N. Li,
    8. F. Wei
    . 2015. Regulation of IL-10 and IL-12 production and function in macrophages and dendritic cells. F1000 Res. 4: F1000 Faculty Rev-1465.
  41. ↵
    1. Thomas, A. C.,
    2. J. T. Mattila
    . 2014. “Of mice and men”: arginine metabolism in macrophages. Front. Immunol. 5: 479.
    OpenUrlCrossRefPubMed
  42. ↵
    1. Agoro, R.,
    2. M. Taleb,
    3. V. F. J. Quesniaux,
    4. C. Mura
    . 2018. Cell iron status influences macrophage polarization. PLoS One 13: e0196921.
    1. Jablonski, K. A.,
    2. S. A. Amici,
    3. L. M. Webb,
    4. J. D. Ruiz-Rosado,
    5. P. G. Popovich,
    6. S. Partida-Sanchez,
    7. M. Guerau-de-Arellano
    . 2015. Novel markers to delineate murine M1 and M2 macrophages. PLoS One 10: e0145342.
  44. ↵
    1. Ying, W.,
    2. P. S. Cheruku,
    3. F. W. Bazer,
    4. S. H. Safe,
    5. B. Zhou
    . 2013. Investigation of macrophage polarization using bone marrow derived macrophages. J. Vis. Exp. DOI: 10.3791/50323.
  45. ↵
    1. Alleva, D. G.,
    2. E. B. Johnson,
    3. F. M. Lio,
    4. S. A. Boehme,
    5. P. J. Conlon,
    6. P. D. Crowe
    . 2002. Regulation of murine macrophage proinflammatory and anti-inflammatory cytokines by ligands for peroxisome proliferator-activated receptor-gamma: counter-regulatory activity by IFN-gamma. J. Leukoc. Biol. 71: 677–685.
    OpenUrlPubMed
  46. ↵
    1. Shen, C. H.,
    2. J. Stavnezer
    . 1998. Interaction of stat6 and NF-kappaB: direct association and synergistic activation of interleukin-4-induced transcription. Mol. Cell. Biol. 18: 3395–3404.
    OpenUrlAbstract/FREE Full Text
  47. ↵
    1. Ma, Y.,
    2. B. Wang,
    3. F. Jiang,
    4. D. Wang,
    5. H. Liu,
    6. Y. Yan,
    7. H. Dong,
    8. F. Wang,
    9. B. Gong,
    10. Y. Zhu, et al
    . 2013. A feedback loop consisting of microRNA 23a/27a and the β-like globin suppressors KLF3 and SP1 regulates globin gene expression. Mol. Cell. Biol. 33: 3994–4007.
    OpenUrlAbstract/FREE Full Text
  48. ↵
    1. Wimmer, N.,
    2. B. Huber,
    3. N. Barabas,
    4. J. Röhrl,
    5. K. Pfeffer,
    6. T. Hehlgans
    . 2012. Lymphotoxin β receptor activation on macrophages induces cross-tolerance to TLR4 and TLR9 ligands. J. Immunol. 188: 3426–3433.
    OpenUrlAbstract/FREE Full Text
  49. ↵
    1. Li, L.,
    2. D. S. Ng,
    3. W. C. Mah,
    4. F. F. Almeida,
    5. S. A. Rahmat,
    6. V. K. Rao,
    7. S. C. Leow,
    8. F. Laudisi,
    9. M. T. Peh,
    10. A. M. Goh, et al
    . 2015. A unique role for p53 in the regulation of M2 macrophage polarization. Cell Death Differ. 22: 1081–1093.
    OpenUrlCrossRefPubMed
  50. ↵
    1. Ma, W. T.,
    2. F. Gao,
    3. K. Gu,
    4. D. K. Chen
    . 2019. The role of monocytes and macrophages in autoimmune diseases: a comprehensive review. Front. Immunol. 10: 1140.
    OpenUrl
  51. ↵
    1. Pathria, P.,
    2. T. L. Louis,
    3. J. A. Varner
    . 2019. Targeting tumor-associated macrophages in cancer. Trends Immunol. 40: 310–327.
    OpenUrlPubMed
  52. ↵
    1. Shapouri-Moghaddam, A.,
    2. S. Mohammadian,
    3. H. Vazini,
    4. M. Taghadosi,
    5. S. A. Esmaeili,
    6. F. Mardani,
    7. B. Seifi,
    8. A. Mohammadi,
    9. J. T. Afshari,
    10. A. Sahebkar
    . 2018. Macrophage plasticity, polarization, and function in health and disease. J. Cell. Physiol. 233: 6425–6440.
    OpenUrlCrossRefPubMed
  53. ↵
    1. Goswami, K. K.,
    2. T. Ghosh,
    3. S. Ghosh,
    4. M. Sarkar,
    5. A. Bose,
    6. R. Baral
    . 2017. Tumor promoting role of anti-tumor macrophages in tumor microenvironment. Cell. Immunol. 316: 1–10.
    OpenUrl
  54. ↵
    1. Hagemann, T.,
    2. J. Wilson,
    3. F. Burke,
    4. H. Kulbe,
    5. N. F. Li,
    6. A. Plüddemann,
    7. K. Charles,
    8. S. Gordon,
    9. F. R. Balkwill
    . 2006. Ovarian cancer cells polarize macrophages toward a tumor-associated phenotype. J. Immunol. 176: 5023–5032.
    OpenUrlAbstract/FREE Full Text
  55. ↵
    1. Hagemann, T.,
    2. T. Lawrence,
    3. I. McNeish,
    4. K. A. Charles,
    5. H. Kulbe,
    6. R. G. Thompson,
    7. S. C. Robinson,
    8. F. R. Balkwill
    . 2008. “Re-educating” tumor-associated macrophages by targeting NF-kappaB. J. Exp. Med. 205: 1261–1268.
    OpenUrlAbstract/FREE Full Text
  56. ↵
    1. Baig, M. S.,
    2. A. Roy,
    3. S. Rajpoot,
    4. D. Liu,
    5. R. Savai,
    6. S. Banerjee,
    7. M. Kawada,
    8. S. M. Faisal,
    9. R. Saluja,
    10. U. Saqib, et al
    . 2020. Tumor-derived exosomes in the regulation of macrophage polarization. Inflamm. Res. 69: 435–451.
    OpenUrl
    1. Komohara, Y.,
    2. K. Ohnishi,
    3. J. Kuratsu,
    4. M. Takeya
    . 2008. Possible involvement of the M2 anti-inflammatory macrophage phenotype in growth of human gliomas. J. Pathol. 216: 15–24.
    OpenUrlCrossRefPubMed
    1. Pritchard, A.,
    2. S. Tousif,
    3. Y. Wang,
    4. K. Hough,
    5. S. Khan,
    6. J. Strenkowski,
    7. B. K. Chacko,
    8. V. M. Darley-Usmar,
    9. J. S. Deshane
    . 2020. Lung tumor cell-derived exosomes promote M2 macrophage polarization. Cells 9: 1303.
    OpenUrl
  59. ↵
    1. Yang, Y.,
    2. J. Qin,
    3. L. Lan,
    4. N. Li,
    5. C. Wang,
    6. P. He,
    7. F. Liu,
    8. H. Ni,
    9. Y. Wang
    . 2014. M-CSF cooperating with NFκB induces macrophage transformation from M1 to M2 by upregulating c-Jun. Cancer Biol. Ther. 15: 99–107.
    OpenUrlCrossRef
  60. ↵
    1. Chaudhuri, A. A.,
    2. A. Y. So,
    3. A. Mehta,
    4. A. Minisandram,
    5. N. Sinha,
    6. V. D. Jonsson,
    7. D. S. Rao,
    8. R. M. O’Connell,
    9. D. Baltimore
    . 2012. Oncomir miR-125b regulates hematopoiesis by targeting the gene Lin28A. Proc. Natl. Acad. Sci. USA 109: 4233–4238.
    OpenUrlAbstract/FREE Full Text
  61. ↵
    1. Cho, H. J.,
    2. J. I. Jung,
    3. D. Y. Lim,
    4. G. T. Kwon,
    5. S. Her,
    6. J. H. Park,
    7. J. H. Park
    . 2012. Bone marrow-derived, alternatively activated macrophages enhance solid tumor growth and lung metastasis of mammary carcinoma cells in a Balb/C mouse orthotopic model. Breast Cancer Res. 14: R81.
    OpenUrlCrossRefPubMed
    1. Sun, Y.,
    2. H. F. Lodish
    . 2010. Adiponectin deficiency promotes tumor growth in mice by reducing macrophage infiltration. PLoS One 5: e11987.
  63. ↵
    1. Xi, J.,
    2. Q. Huang,
    3. L. Wang,
    4. X. Ma,
    5. Q. Deng,
    6. M. Kumar,
    7. Z. Zhou,
    8. L. Li,
    9. Z. Zeng,
    10. K. H. Young, et al
    . 2018. miR-21 depletion in macrophages promotes tumoricidal polarization and enhances PD-1 immunotherapy. Oncogene 37: 3151–3165.
    OpenUrl
  64. ↵
    1. Tan, B.,
    2. X. Shi,
    3. J. Zhang,
    4. J. Qin,
    5. N. Zhang,
    6. H. Ren,
    7. M. Qian,
    8. S. Siwko,
    9. K. Carmon,
    10. Q. Liu, et al
    . 2018. Inhibition of rspo-lgr4 facilitates checkpoint blockade therapy by switching macrophage polarization. Cancer Res. 78: 4929–4942.
    OpenUrlAbstract/FREE Full Text
  65. ↵
    1. Malyshev, I.,
    2. Y. Malyshev
    . 2015. Current concept and update of the macrophage plasticity concept: intracellular mechanisms of reprogramming and M3 macrophage “switch” phenotype. BioMed Res. Int. 2015: 341308.
  66. ↵
    1. Chung, S.,
    2. R. Ranjan,
    3. Y. G. Lee,
    4. G. Y. Park,
    5. M. Karpurapu,
    6. J. Deng,
    7. L. Xiao,
    8. J. Y. Kim,
    9. T. G. Unterman,
    10. J. W. Christman
    . 2015. Distinct role of FoxO1 in M-CSF- and GM-CSF-differentiated macrophages contributes LPS-mediated IL-10: implication in hyperglycemia. J. Leukoc. Biol. 97: 327–339.
    OpenUrlCrossRefPubMed
  67. ↵
    1. Lee, E. G.,
    2. D. L. Boone,
    3. S. Chai,
    4. S. L. Libby,
    5. M. Chien,
    6. J. P. Lodolce,
    7. A. Ma
    . 2000. Failure to regulate TNF-induced NF-kappaB and cell death responses in A20-deficient mice. Science 289: 2350–2354.
    OpenUrlAbstract/FREE Full Text
  68. ↵
    1. Nanbakhsh, A.,
    2. A. Srinivasamani,
    3. S. Holzhauer,
    4. M. J. Riese,
    5. Y. Zheng,
    6. D. Wang,
    7. R. Burns,
    8. M. H. Reimer,
    9. S. Rao,
    10. A. Lemke, et al
    . 2019. Mirc11 disrupts inflammatory but not cytotoxic responses of NK cells. Cancer Immunol. Res. 7: 1647–1662.
    OpenUrlAbstract/FREE Full Text
  69. ↵
    1. Genard, G.,
    2. S. Lucas,
    3. C. Michiels
    . 2017. Reprogramming of tumor-associated macrophages with anticancer therapies: radiotherapy versus chemo- and immunotherapies. Front. Immunol. 8: 828.
    OpenUrlPubMed
  70. ↵
    1. Kouri, F. M.,
    2. C. Ritner,
    3. A. H. Stegh
    . 2015. miRNA-182 and the regulation of the glioblastoma phenotype - toward miRNA-based precision therapeutics. Cell Cycle 14: 3794–3800.
    OpenUrl
  71. ↵
    1. Bhaskaran, V.,
    2. M. O. Nowicki,
    3. M. Idriss,
    4. M. A. Jimenez,
    5. G. Lugli,
    6. J. L. Hayes,
    7. A. B. Mahmoud,
    8. R. E. Zane,
    9. C. Passaro,
    10. K. L. Ligon, et al
    . 2019. The functional synergism of microRNA clustering provides therapeutically relevant epigenetic interference in glioblastoma. Nat. Commun. 10: 442.
    OpenUrl
  72. ↵
    1. Parayath, N. N.,
    2. A. Parikh,
    3. M. M. Amiji
    . 2018. Repolarization of tumor-associated macrophages in a genetically engineered nonsmall cell lung cancer model by intraperitoneal administration of hyaluronic acid-based nanoparticles encapsulating microRNA-125b. Nano Lett. 18: 3571–3579.
    OpenUrl
    1. Su, M. J.,
    2. H. Aldawsari,
    3. M. Amiji
    . 2016. Pancreatic cancer cell exosome-mediated macrophage reprogramming and the role of MicroRNAs 155 and 125b2 transfection using nanoparticle delivery systems. Sci. Rep. 6: 30110.
    OpenUrl
  74. ↵
    1. Tran, T. H.,
    2. S. Krishnan,
    3. M. M. Amiji
    . 2016. MicroRNA-223 induced repolarization of peritoneal macrophages using CD44 targeting hyaluronic acid nanoparticles for anti-inflammatory effects. PLoS One 11: e0152024.
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section