A double feedback loop mediated by microRNA-23a/27a/24-2 regulates M1 versus M2 macrophage polarization and thus regulates cancer progression

Abstract

In response to microenvironmental signals, macrophages undergo different types of activation, including the "classic" pro-inflammatory phenotype (also called M1) and the "alternative" anti-inflammatory phenotype (also called M2). Macrophage polarized activation has profound effects on immune and inflammatory responses, but mechanisms underlying the various types of macrophage is still in its infancy. In this study, we reported that M1-type stimulation could down-regulate miR-23a/27a/24-2 cluster transcription through the binding of NF-κB to this cluster's promoter and that miR-23a in turn activated the NF-κB pathway by targeting A20 and thus promoted the production of pro-inflammatory cytokines. Furthermore, STAT6 occupied the miR-23a/27a/24-2 cluster promoter and activated their transcription in IL-4-stimulated macrophages. In addition, miR-23a in turn suppressed the JAK1/STAT-6 pathway and reduced the production of M2 type cytokines by targeting JAK1 and STAT-6 directly, while miR-27a showed the same phenotype by targeting IRF4 and PPAR-γ. The miR-23a/27a/24-2 cluster was shown to be significantly decreased in TAMs of breast cancer patients, and macrophages overexpressing the miR-23a/27a/24-2 cluster inhibited tumor growth in vivo. Taken together, these data integrated microRNA expression and function into macrophage polarization networks and identified a double feedback loop consisting of the miR-23a/27a/24-2 cluster and the key regulators of the M1 and M2 macrophage polarization pathway. Moreover, miR-23a/27a/24-2 regulates the polarization of tumor-associated macrophages and thus promotes cancer progression.

Keywords: macrophage polarization; miRNA; tumor associated macrophage; tumor microenvironment; tumor therapy.

Figure 1. The miR-23a/27a/24-2 cluster was simultaneously down-regulated by M1-type stimuli and up-regulated by M2-type stimuli

A. qRT-PCR analysis of the relative expression of the miRNAs in BMDMs treated with PBS (M0), 1 μg/ml LPS plus 20 ng/ml IFN-γ (M1), or 100 ng/ml IL-4 (M2). B. qPCR analysis of the relative expression of the mature miRNAs after the PM cells were stimulated with 1 μg/ml LPS for 0.5 h, 3 h, 6 h and 12 h. C. qPCR analysis of the relative expression of the cluster precursor after the PM cells were stimulated with 1 μg/ml LPS for 3 h and 12 h. D. qPCR analysis of the relative expression of the cluster precursor after the PM cells were stimulated with100 ng/ml IL-4 for 12 h and 24 h. E. qPCR analysis of the relative expression of the mature miRNAs after the PM cells were stimulated with 100 ng/ml IL-4 for 24 h. F. qPCR analysis of the relative expression of the cluster precursor after the RAW264.7 cells were stimulated with 1 μg/ml LPS, 20 ng/ml GM-CSF, 100 ng/ml IL-4, 60 ng/ml IL-13, or 20 ng/ml M-CSF. G. qPCR analysis of the relative expression of the mature miRNAs after the RAW264.7 cells were stimulated with 1 μg/ml LPS, 20 ng/ml GM-CSF, 100 ng/ml IL-4, 60 ng/ml IL-13, or 20 ng/ml M-CSF. H. qPCR analysis of the relative expression of the cluster precursor after the BMDMs were stimulated with 1 μg/ml LPS, 20 ng/ml GM-CSF, 100 ng/ml IL-4, 60 ng/ml IL-13, or 20 ng/ml M-CSF. I. qPCR analysis of the relative expression of the mature miRNAs after the BMDMs were stimulated with 1 μg/ml LPS, 20 ng/ml GM-CSF, 100 ng/ml IL-4, 60 ng/ml IL-13, or 20 ng/ml M-CSF. Mean±SD were obtained from three independent experiments. *, p < 0.05; **, p < 0.01.

Figure 2. IL-4 promoted transcription of the cluster through STAT6, and LPS negatively regulated transcription of the cluster through NF-κB

A. The miR-23a/27a/24-2 cluster was located on chromosome 8 (84208518∼ 84208921) in the genome of mus musculus (Upper panel). NF-κB and STAT-X binding sites were predicted to be located within the promoter of the cluster using bioinformatics. B. Wild type and three mutant promoters of the cluster were cloned into the pGL3-basic vector; the binding sites were inactivated by point mutations shown in red. C. Relative firefly luciferase activity derived from WT, NF-κB mut, STAT-X mut, and NF-κB/STAT-X mut constructs following transfection into 293-T cells. All values were normalized to renilla luciferase activity produced from a co-transfected control plasmid. Error bars represent standard deviations from 6 independent transfections. D. ChIP-qPCR assay with two different primers documenting that the TF that bound to the cluster promoter is STAT6. E. qPCR analysis of the cluster precursor expression in RAW264.7 cells transfected with the pwpxl-STAT6 construct or STAT6 siRNA, following stimulation with 50 ng/ml IL-4 or PBS. F. qPCR analysis of the relative expression of the cluster precursor after the RAW264.7 cells were stimulated with another NF-κB pathway activator, CpG, for 0 h, 0.5 h, 3 h, 6 h and 12 h. G. qPCR analysis of the relative expression of the three mature miRNAs after the RAW264.7 cells were stimulated with another NF-κB pathway activator, CpG OND, for 0 h, 0.5 h, 3 h, 6 h and 12 h. H. qPCR analysis of the relative expression of the three mature miRNAs after the RAW264.7 cells were stimulated with another NF-κB pathway activator, Poly(I:C), for 0 h, 0.5 h, 3 h, 6 h and 12 h. I. Western blots documenting the total IκBα and p-IκBα levels in RAW264.7 cells after stimulation with the NF-κB pathway activators, CpG OND, LPS, and Poly (I: C), normalized to GAPDH levels. Mean±SD were obtained from three independent experiments. *, p < 0.05; **, p < 0.01.

Figure 3. MiR-23a/27a/24-2 promoted the expression of pro-inflammatory cytokines

Q-PCR analysis of the relative expression of IL-1β, IL-6, and TNF-α mRNA after RAW264.7 cells were transfected with miR-23a mimics A., miR-27a mimics B., or miR-24-2 mimics C., for 42 h and stimulated with LPS for 6 h. D. qPCR analysis of the relative expression or IL-1β, IL-6, and TNF-α mRNA after PMs were transfected with miR-23a mimics or inhibitors for 42 h and stimulated with LPS for 6 h. qPCR analysis of the relative expression of Arg1, Fizz1, and IL-10 mRNA after PMs were transfected with miR-23a mimics E., miR-23a inhibitors F., miR-27a mimics G., miR-27a inhibitors H., miR-24-2 mimics I., or miR-24-2 inhibitors J. and stimulated with IL-4. Mean±SD were obtained from three independent experiments. *, p < 0.05; **, p < 0.01.

Figure 4. MiR-23a promoted M1 polarization by targeting A20

A. Relative firefly luciferase activity derived from the pGL3-p65 promoter co-transfected into 293-T cells with the TK plasmids and the miR-23a or control mimics. All values were normalized to renilla luciferase activity produced from a co-transfected control plasmid. Error bars represent standard deviations from 6 independent transfections. B. MiR-23a was predicted by bioinformatics to bind to the 3′UTR of A20. C. Relative firefly luciferase activity derived from A20-3′UTR and A20-3′UTR-MUT co-transfected into 293-T cells with the TK plasmids and the miR-23a or control mimics. All values were normalized to renilla luciferase activity produced from a co-transfected control plasmid. Error bars represent standard deviations from 3 independent transfections. The A20 protein levels D. and IL-6 levels in media E. of RAW264.7 cells transfected with pcDNA3.1-A20 for 24 hours followed by miR-23a mimics transfection. F. Western blots documenting the expression of A20 with PBS or 0.1 mg/ml LPS stimulation for 3 h, 6 h, 12 h in PMs, normalized to GAPDH levels. G. Western blots documenting the expression of A20 after transfection with miR-23a or control mimics in PMs, normalized to GAPDH levels.

Figure 5. MiR-23a and miR-27a inhibited M2 polarization by targeting JAK1/STAT6 and IRF4/PPAR-γ

MiR-23a was predicted by bioinformatics to bind to the 3′UTR of JAK1 at two different sites A. and the 3′UTR of STAT6 B. MiR-27a was predicted by bioinformatics to bind to the 3′UTR of IRF4 C. and PPAR-γ D. Relative firefly luciferase activity derived from STAT6-3′UTR and STAT6-3′UTR-MUT E., or JAK1-3′UTR, JAK1-3′UTR-MUT1, JAK1-3′UTR-MUT2, and JAK1-3′UTR-MUT F. co-transfected into 293-T cells with the TK plasmids and the miR-23a or control mimics. Relative firefly luciferase activity derived from IRF4-3′UTR and IRF4-3′UTR-MUT G., or PPAR-γ-3′UTR and PPAR-γ-3′UTR-MUT H. co-transfected into 293-T cells with the TK plasmids and the miR-27a or control mimics. Error bars represent standard deviations from 3 independent transfections, *p < 0.05, **p < 0.01, ***p < 0.001. I. Western blots documenting the expression of STAT6 and JAK1 after transfection with miR-23a mimics or inhibitors in PMs, normalized to GAPDH levels. J. Western blots documenting the expression of IRF4 and PPAR-γ after transfection with miR-27a mimics or inhibitors in PMs, normalized to GAPDH levels. K. Western blots documenting the expression of STAT6, JAK1, IRF4, and PPAR-γ with PBS or 0.1 mg/ml IL-4 stimulation for 12 h in PMs, normalized to GAPDH levels.

Figure 6. The miR-23a/27a/24-2 cluster was down-regulated in TAMs in breast cancer patients

Expression level of miR-23a A.,miR-27a B., and miR-24-2. C. in TAMs isolated from breast cancer tissue compared with paired PBMCs of stage II or III patients, as revealed by qRT-PCR assay (n = 20). D. The protein levels of NF-κB p65 subunit and STAT6 in PBMCs and TAMs. The expression levels of M1 and M2 cytokines in TAMs after transfected with miR-23a inhibitors E., miR-27a inhibitors F. or miR-24-2 inhibitors G.

Figure 7. The miR-23a/27a/24-2 cluster reduced tumor growth in vivo

Approximately 3:1 4T1 cells and RAW264.7-miR-23a/27a/24-2 cells or control cells were s.c. co-injected into 4-6 week-old BALB/c female mice. 11 days later, tumor lengths and widths were measured with a caliper every 3 or 4 days until day 28 (n = 5). Tumor volumes at different time points were shown in A. and B. Tumor weights and average weights at the experimental end point were shown in C. and D. Approximately 3:1 4T1 cells and BMDM-miR-23a/27a/24-2 antagomir cells or control cells were s.c. co-injected into 4-6 week-old BALB/c female mice. 6 days later, tumor lengths and widths were measured with a caliper every 3 days for 3 weeks (n = 5). Tumor volumes at different time points were shown in E. Tumor weights and average weights at the experimental end point were shown in F. and G. H. Schematic representing M2 polarization of TAMs through the regulation of the miRNA-23a cluster. ***p < 0.001; **p < 0.01; *p < 0.05.

Figure 8. The signal pathway consists of the miR-23a cluster and important regulators during M1 and M2 polarization