MicroRNA-21 Limits In Vivo Immune Response-Mediated Activation of the IL-12/IFN-γ Pathway, Th1 Polarization, and the Severity of Delayed-Type Hypersensitivity

Thelper 1 and Th2 cytokines oppose each other’s function and typically exist in a balanced state (1). An altered balance between Th1 and Th2 cytokines is responsible for a variety of immunoinflammatory disorders. For example, allergic inflammation is typically characterized by a Th2 cell cytokine-like response, with overexpression of Th2 (e.g., IL-4, IL-5, IL-13, and IL-25) and underexpression of Th1 (e.g., IFN-γ) cytokines (2, 3). Although transcriptional regulation of Th cell polarization and function has received considerable attention, the role of microRNAs (miRs) in this process has not been reported, except in an early recent report showing that let-7 regulates IL-13 expression (4) and in an elegant study showing that miR-126 suppresses the effector function of Th2 cells (5).

We have reported previously that allergic airway inflammation, triggered by allergen or IL-13, dysregulates a series of miRs (6). Among these miRs, miR-21 is the most upregulated miR in multiple models of experimental asthma, and we found that the IL-12p35 3′ untranslated region (UTR) contains a functional miR-21 binding site in a heterologous cell line (6). IL-12 is a major cytokine involved in Th1 cell polarization. It is a heterodimeric cytokine composed of a p35 and a p40 subunit, with the heterodimer (IL-12p70) being the bioactive protein (7). Repression of IL-12p35 expression by miR-21 could lead to decreased IL-12p70 production and in part explain the exaggerated Th2 response seen in asthma (8). This exciting possibility, which would represent a new paradigm for controlling polarized adaptive immune responses, remains unproven given that miR-21’s ability to suppress IL-12p35 is limited to analysis in a heterologous cell line following artificial transfection and assessment of an engineered luciferase reporter construct (6). In addition, although miR-21 is strongly upregulated following experimental asthma induction, at least 20 other miRs are dysregulated, which could induce direct or indirect effects on a variety of complementary pathways. However, it is notable that the miR-21 binding site in the IL-12p35 3′ UTR is conserved over a variety of species ranging from humans to platypus (6), supporting the potential of our hypothesis. The role of miR-21 will likely extend to a variety of diseases including malignancies, as it is notable that miR-21 is also consistently elevated across a variety of tumors (9).

In this study, we sought to determine the in vivo impact of miR-21 on murine models of hypersensitivity in the lung and skin using a novel strain of miR-21 gene-targeted mice. We first identified mRNA transcripts dysregulated in the lung during allergic inflammation by loss of miR-21. We then used an unbiased systems biology analysis approach that identified an unexpected prominent dysregulation of IL-12/IFN-γ pathways as the most significantly affected in the lungs of OVA-challenged miR-21^−/− mice compared with miR-21^+/+ littermates. In turn, miR-21–deficient mice had increased IFN-γ and decreased IL-4 levels in the lung compared with wild-type littermates. This was associated with reduced eosinophilia in the lungs of miR-21^−/− mice. To test for the cellular origins of these effects, we next demonstrated that miR-21^−/− dendritic cells produced significantly more IL-12 compared with wild-type dendritic cells after LPS stimulation, potentiated by IFN-γ costimulation. Furthermore, OVA-challenged miR-21^−/− CD4⁺ T lymphocytes produced increased IFN-γ and decreased IL-4 compared with control cells. To broaden our finding, we examined the impact of miR-21 deficiency in an independent, Th1-associated cutaneous delayed-type hypersensitivity model. We demonstrated that the loss of miR-21 significantly enhanced the Th1-associated cutaneous delayed-type hypersensitivity responses. These data demonstrate that miR-21 has a central role in establishing the fine balance of Th1 versus Th2 responses to Ags and suggest that targeting miR-21 and understanding variations in its activity may lead to new treatments and preventions for a variety of diseases that exhibit dysregulated Th1/Th2 balance such as allergic asthma.

The pFlexible-based miR-21 gene-targeting vector (10), modified to permit diphtheria toxin A-negative selection in embryonic stem (ES) cells, contained 6.3 kb 5′ homologous sequence including exon 12 of the Tmem49 (VMP1) gene, 1.4 kb loxP-flanked sequence containing pre–miR-21, and a Flprecombinase recognition site-flanked pgk-puroΔTK gene, followed by 1.5 kb 3′ homologous sequence. Homologous and conditional miR-21 sequence was amplified from CJ7 ES cell-derived genomic DNA by PCR and verified by DNA sequencing. The linearized targeting vector was electroporated into CJ7 ES cells, and PCR-mediated screening of 114 puromycin-resistant cells yielded 31 candidates. Correct homologous recombination in three candidate ES cell clones was confirmed by Southern analysis and locus-specific PCR selectively amplifying the targeted miR-21 allele combined with DNA sequencing. Two independently targeted ES cell clones were injected into C57BL/6-derived blastocysts, and chimeric offspring were bred with C57BL/6 EIIA-Cre mice to delete the pre–miR-21 containing conditional sequence. Germline deletion of the conditional miR-21 allele was demonstrated by genomic PCR in heterozygous mice that had been backcrossed to remove the cre transgene. Heterozygous intercrosses yielded homozygous mice at the expected Mendelian frequency. Mouse lines derived from two independently targeted ES cell clones were phenotypically indistinguishable. The mice were backcrossed for two generations into the C57BL/6 background for the murine asthma models. The mice were backcrossed for five generations into the C57BL/6 background for the delayed-type hypersensitivity experiments. Littermate controls were used for all experiments. All animals used were housed under specific pathogen-free conditions in accordance with institutional guidelines. The Institutional Animal Care and Use Committee of the Cincinnati Children’s Hospital Medical Center approved the use of animals in these experiments.

Total RNA from wild-type miR-21^+/+, heterozygous conditional miR-21^f/+, and homozygous gene-targeted miR-21^−/− CJ7 ES cells was extracted using standard procedures. For miRNA expression analysis, 20 μg total RNA per sample was electrophoretically separated on a 15% polyacrylamide Tris-borate-EDTA/urea gel and transferred to a nylon membrane via semidry transfer. Expression of miR-21, miR-294, and 5S RNA was detected by sequentially hybridizing the membrane with radioactively labeled DNA oligonucleotide probes. For mRNA expression analysis, 30 μg denatured total RNA per sample was electrophoretically separated on a 1.2% (w/v) agarose gel and transferred to a nylon membrane via alkaline downward capillary transfer. Expression of Tmem49 (VMP1) and GAPDH mRNA was detected by sequential hybridization of the membrane with radioactively labeled DNA probes.

Experimental asthma was induced by injection with 100 μg OVA and 1 mg aluminum hydroxide as adjuvant on day 0 and day 10, followed by 50 μg OVA or saline intranasal challenges on days 20, 21, and 22. Mice were sacrificed 18–24 h after the last challenge (11).

Bronchoalveolar lavage fluid (BALF) was collected according to previously published methods (12). Total cell numbers were counted with a hemacytometer. Cytospin preparations of 7.5 × 10⁴ cells were stained with Diff-Quik solution, and differential cell counts were determined using standard morphological criteria. At least 250 cells were counted per slide for differential cell counts. Cytokine levels of IFN-γ, IL-4, and CXCL9 were determined using the Milliplex Multi-Analyte Profiling Immunoassay (Millipore). Levels of CCL17 were determined using DuoSet ELISA kits (R&D Systems) according to the manufacturer’s protocol.

Quantification of eosinophil infiltration in the lung tissue was performed as previously described (12). Briefly, lung tissue eosinophils were identified by anti-major basic protein staining. Quantification of immunoreactive cells was performed by counting the positive-stained cells under low-power magnification, and eosinophil levels were expressed as the number of eosinophils per square millimeter.

Total RNA was isolated using the miRNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol. Levels of miR expression were measured quantitatively by using the TaqMan MicroRNA Assay (Applied Biosystems) following the manufacturer’s protocol and assayed on the Applied Biosystems 7900HT Real-Time PCR System (Applied Biosystems) (13). Normalization was performed using U6 small nuclear RNA. Relative expression was calculated using the comparative threshold cycle method as previously described (14).

Total RNA was reverse transcribed using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems). All primer/probe sets were obtained from Applied Biosytems. Samples were analyzed by TaqMan quantitative real-time PCR (qRT-PCR) for IL-12p35 (assay ID: Mm00434169_m1), IFN-γ (assay ID: Mm_00801778_m1), and CXCL10 (assay ID: Mm_00445235_m1) and normalized to HPRT1 (assay ID: Mm00446968_m1). Relative expression was calculated using the comparative threshold cycle method (14).

Bone marrow-derived dendritic cells were prepared by culturing cells in RPMI 1640 medium supplemented with 40 ng/ml GM-CSF, 10 ng/ml IL-4, 10% FBS, 1% penicillin/streptomycin, and 50 μM 2-ME for 7 d as previously described (15).

Bone marrow-derived dendritic cells were stimulated with 1 μg/ml LPS (strain: 055:B5; Sigma-Aldrich) or 1 μg/ml LPS plus 100 IU/ml IFN-γ (eBioscience) for 48 h. IL-12p70 concentrations from cell-culture supernatants were measured using the Legend Max Mouse IL-12p70 ELISA Kit (BioLegend) according to the manufacturer’s protocol. TNF-α and IL-6 concentrations from cell-culture supernatants were measured using ELISA Max Deluxe kits (BioLegend) according to the manufacturer’s protocol. To measure the levels of IL-23 production, bone marrow-derived dendritic cells were stimulated with 1 μg/ml LPS for 12, 24, or 48 h. IL-23 concentrations from cell-culture supernatants were measured using IL-23 DuoSet ELISA kit (R&D Systems) according to manufacturer’s protocol. To measure the levels of IFN-γ production, bone marrow-derived dendritic cells were stimulated with 10 ng/ml IL-12, 10 ng/ml IL-23, or 1 μg/ml LPS for 48 h. IFN-γ concentrations from cell-culture supernatants were measured using IFN-γ ELISA Ready-Set-Go kit (eBioscience) according to the manufacturer’s protocol.

Total RNA was isolated from lung tissue using the miRNeasy Mini Kit according to the manufacturer’s protocol (Qiagen). RNA quality was assessed by using the Agilent 2100 bioanalyzer (Agilent Technologies), and only samples with an RNA integrity number >8 were used. Microarray analysis was performed as previously described (11). The Affymetrix mouse Gene 1.0ST array was used (Affymetrix), and data were analyzed using GeneSpring software (Agilent Technologies). Global scaling was performed to compare genes from chip to chip, and a base set of probes was generated by requiring a minimum raw expression level of 20th percentile out of all probes on the microarray. To identify miR-21–regulated genes, the probe sets were then baseline transformed and filtered on at least a 1.2-fold change between miR-21^+/+ and miR-21^−/− mice in either saline- or OVA-challenged mice. A cutoff of 1.2-fold was selected because miRs are currently reported to have only mild to moderate effects on gene expression, with some targets repressed without detectable change in mRNA levels (16, 17). Statistical significance was determined at p < 0.05 with Benjamini Hochberg false discovery rate correction. The resulting list of genes was clustered using hierarchical clustering and a heat map was generated. Pathway analysis of differentially regulated genes was carried out using Ingenuity Pathway Analysis (Ingenuity Systems) and Toppgene/Toppcluster (18). The microarray data have been deposited into the Array Express database (http://www.ebi.ac.uk/arrayexpress) with accession number E-MEXP-3119 in compliance with minimum information about microarray experiment standards.

Global human transcriptome analysis of miR-21 expression patterns were carried out using two large human Affymetrix gene expression datasets (Affymetrix), one based on a series of >2000 human cancers (the Cancer ExPO database, National Center for Biotechnology Information GSE2109) and the other the Human Body Index, GSE7307, which contains 677 samples representing >90 human tissue samples and isolated cell types. CEL files were subjected to robust multichip average normalization and then transformed into a normalized matrix representing the relative expression of each probe across each of the two large datasets. The data were then analyzed by Pearson correlation analyses to identify probesets on the HG U-133 plus 2.0 gene chip that were regulated most similarly to the Affymetrix probeset 224917_at (Affymetrix) that we had predicted would be the best indicator of miR-21 host gene RNA expression. Functional enrichment analysis of the gene list was carried out using the Toppgene suite (19).

Splenic CD4⁺ T cells were isolated using CD4⁺ T cell isolation kit II (Miltenyi Biotec) according to the manufacturer’s protocol. The CD4⁺ T cells were then restimulated with Dynabeads Mouse T-Activator CD3/CD28 (Invitrogen) according to the manufacturer’s protocol. Supernatants were collected after 72 h, and concentrations of IFN-γ and IL-4 were measured using ELISA Max deluxe kits (BioLegend) according to the manufacturer’s protocol.

Isolation and restimulation of total splenocytes were performed as previously described (20). Briefly, spleens were aseptically removed from mice. Single-cell suspensions were obtained by mincing the spleen and gently pressing the fragments through a 70-μm filter. The cell suspensions were washed twice in Advanced RPMI 1640 medium supplemented with 10% FBS, 2 mM l-glutamine, 1 mM sodium pyruvate, 50 U/ml penicillin, and 50 μg/ml streptomycin. After counting by trypan blue exclusion, they were adjusted to a density of 1 × 10⁷ cells/ml. Cells were plated at 1 ml/well in 24-well cell-culture plates and challenged with OVA at a final concentration of 100 μg/ml or with an equivalent volume of saline. Supernatants were collected after 72 h, and concentrations of IFN-γ and IL-4 were measured using ELISA Max deluxe kits (BioLegend) according to the manufacturer’s protocol.

Splenic NK cells were isolated using NK Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s protocol. The NK cells were then stimulated with 10 ng/ml IL-12. Supernatants were collected after 48 h, and concentrations of IFN-γ were measured using ELISA Max deluxe kits (BioLegend) according to the manufacturer’s protocol.

4-hydroxy-3-nitrophenylacetyl-hydroxysuccinimide ester (NP-OSu)-mediated delayed-type hypersensitivity was induced by immunizing 9–12-wk-old miR-21^+/+ and miR-21^−/− mice s.c. in two sites, one on each ventral flank, with 50 μl 2.5% NP-OSu (Biosearch Technologies) dissolved in DMSO, followed by s.c. injection of 100 μl 1× borate-buffered saline (pH 8.6) in the dorsal midline. Six days later, the left rear footpad was challenged with 25 μl 2.5% NP-OSu freshly diluted 1:20 in PBS (pH 7.8). Control mice were sensitized with DMSO and challenged with NP-OSu solution. The left and right footpad thickness was quantified 24 h after the challenge by using a digital micrometer with 0.01-mm resolution (Mitutoyo). After euthanasia, the left and right rear paws were removed uniformly at the ankle joint and weighed. NP-OSu–induced increases in footpad thickness or weight were calculated by subtracting the right footpad thickness or weight from that of the left footpad.

Statistical analyses were performed with Student t test or one-way ANOVA with Tukey post hoc test where appropriate. Statistical significance and p values were indicated on the figures where appropriate. The p values <0.05 were considered statistically significant.

Targeted disruption of the miR-21 coding sequence was carried out as shown in Fig. 1A. Deletion of miR-21 coding sequence was confirmed by Northern blot and PCR analyses (Fig. 1B, 1C). Regulatory sequences of miR-21 were located adjacent to the Tmem49 gene (also known as VMP1). Tmem49 mRNA expression was unaltered in miR-21^−/− mice (Fig. 1D). miR-21^−/− mice appeared indistinguishable from their wild-type and heterozygous littermates and were delivered in expected Mendelian ratios (n > 100), as recently reported in an independent miR-21^−/− mouse strain (21).

To identify genes for which regulation was affected by the absence or presence of miR-21 in normal or immune-challenged conditions, we performed gene expression microarray analyses of RNA from lungs of saline- and OVA-challenged miR-21^+/+ and miR-21^−/− mice. At baseline, using a statistical cutoff of p < 0.05 and >1.2-fold difference, there were 54 gene-associated transcripts that exhibited differential expression between miR-21^+/+ mice and miR-21^−/− mice (Supplemental Table I). Following OVA challenge, miR-21^+/+ and miR-21^−/− mice both exhibited robust induction of allergen-associated gene signatures. Post hoc filtering to identify allergen response transcripts most significantly affected by the loss of miR-21 identified 316 probe sets that differed by >1.2-fold at p < 0.05 (Supplemental Table II). These included 37 of the 54 saline baseline-affected probes that were also similarly affected under OVA challenge conditions. Hierarchical cluster analysis of the differentially expressed genes identified four striking clusters of differentially expressed genes between miR-21^+/+ and miR-21^−/− mice (Fig. 2A): miR-21^−/− hyperactivated, with notable enrichment for IFN-γ signature genes; miR-21^−/− deactivated and miR-21^−/− hyperrepressed, which include Th2-associated genes such as the chemokine CCL17 (22); and miR-21^−/− derepressed, including IL-18, which is associated with Th1 and Th2 response (Fig. 2A) (23, 24). A full list of genes in these four clusters can be found in Supplemental Table II.

Genes that were differentially regulated by miR-21 were then analyzed using Ingenuity Pathway Analysis. The most significant alteration of all pathways was upregulation of IL-12– and IFN-γ–dependent networks in miR-21^−/− mice compared with miR-21^+/+ littermates (Fig. 2B). The individual nodes in the pathway including genes and/or gene categories identified by the Ingenuity Pathway Analysis are listed in Supplemental Table III. To better understand the potential molecular basis and consequences of the dysregulated IL-12/IFN-γ pathway, we performed biological network analyses using Toppcluster, a multiple gene list feature analyzer for comparative enrichment clustering and network-based dissection of biological systems (Fig. 2C) (18). Unlike the Ingenuity Pathway Analysis approach that is focused on gene and protein interactions, Toppcluster enables connections to be made between genes and gene-associated features that encompass the positive functions of genes (e.g., participation in a pathway, biological process, and protein–protein interactions), regulatory mechanisms via transcription factors or miRNAs, and the consequences of loss of gene functions as evidenced by mouse or human disease phenotypes. This approach provided further insight into relationships among the miR-21^−/− dysregulated genes relative to IFN-γ signaling and responses and highlighted the extensive upregulation of genes that are either positive activators of IFN-γ signaling or its downstream effectors. For example, targeted gene disruption of several of these overexpressed genes give phenotypes in mice that cause decreased IFN-γ production and/or increased IL-4 expression (Fig. 2C).

In conjunction with our prior demonstration that miR-21 has the capacity to directly target the IL-12p35 3′ UTR based on a luciferase reporter assay in heterologous cell in vitro (6), these results and global systems biological analyses demonstrate that in the midst of a complex large-scale allergic airway hypersensitivity immune response, miR-21 exerts a strong polarizing role in modifying the effects of cardinal immune activation pathways.

To test the impact of miR-21 on the levels of Th1 and Th2 cytokines generated by the lung in response to allergen challenge, BALF of miR-21–deficient mice were compared with wild-type littermate controls. Consistent with the results from our systems biology analysis, miR-21^−/− mice had increased levels of the Th1 cytokine IFN-γ as well as CXCL9, a chemokine known to be induced by IFN-γ (Fig. 3A, 3B) (25), and decreased levels of the Th2-associated cytokines IL-4 and CCL17 (Fig. 3C, 3D) (26). We also measured the levels of IL-12p35 by qRT-PCR and found increased levels of IL-12p35 in the miR-21–deficient mice compared with wild-type littermate controls (Fig. 3E).

Based on the strong downregulation of multiple eosinophil-specific transcripts and Th2-associated chemokines in the microarray data, we questioned whether the loss of miR-21 caused loss of eosinophils in the lung. miR-21^−/− mice had significantly reduced levels of eosinophil infiltration in the BALF after allergen challenge compared with wild-type littermate controls (Fig. 4A), with correspondingly reduced eosinophil percentages in the BALF (Fig. 4B) and reduced eosinophil level in the lung tissue (Fig. 4C). These results indicate that the increased IFN-γ polarization had functional and specific consequences, as the levels of neutrophils, lymphocytes, and macrophages were not significantly changed in the BALF (Fig. 4A).

To find a mechanistic basis of the increased Th1 response in the miR-21–deficient mice after allergen challenge, we cultured bone marrow-derived dendritic cells from miR-21^+/+ and miR-21^−/− mice and stimulated the dendritic cells with LPS alone or LPS plus IFN-γ. We then measured the levels of IL-12p70 in the supernatant. Compared to dendritic cells from wild-type littermates, dendritic cells from miR-21–deficient mice produced significantly increased levels of IL-12p70 after LPS or LPS/IFN-γ stimulation (Fig. 5A). By contrast, there was no change in the levels of TNF-α or IL-6 produced, demonstrating specificity of the increased IL-12p70 production by miR-21–deficient dendritic cells (Fig. 5B, 5C). Because IL-23 production from dendritic cells is also an important regulator of Th1 versus Th2 balance, we measured the levels of IL-23 production from bone marrow-derived dendritic cells after 12, 24, or 48 h of LPS stimulation. The bone marrow-derived dendritic cells produced maximum amounts of IL-23 after 12 h of LPS stimulation, consistent with previous reports in the literature (27). However, compared with dendritic cells from wild-type littermates, dendritic cells from miR-21–deficient mice produced similar levels of IL-23 after 12, 24, or 48 h of LPS stimulation (Fig. 5D).

To further support the hypothesis that miR-21 regulates Th1 versus Th2 responses, we compared cytokine responses in isolated CD4⁺ T cells from spleens of OVA-challenged miR-21^−/− mice and wild-type controls. Compared to the wild-type littermate controls, anti-CD3/CD28–restimulated CD4⁺ T cells from miR-21^−/− mice produced increased IFN-γ and decreased IL-4, with a correspondingly increased IFN-γ/IL-4 ratio (Fig. 6A–C). Restimulation of total splenocytes from OVA-challenged mice with OVA showed similar results (Fig. 6D, 6E). By contrast, anti-CD3/CD28–restimulated CD4⁺ T cells from naive miR-21^−/− mice produced similar levels of IFN-γ and IL-4 as the naive wild-type littermates (Fig. 6F, 6G).

Because NK cells and dendritic cells are also important IFN-γ–producing cells, we measured the levels of IFN-γ production by IL-12–stimulated NK cells from spleen and IL-12–, IL-23–, or LPS-stimulated bone marrow-derived dendritic cells. IL-12–stimulated NK cells from miR-21^−/− mice produced similar amounts of IFN-γ as NK cells from wild-type littermates (Fig. 7A). IL-12–stimulated dendritic cells from miR-21^−/− mice also produced similar amounts IFN-γ as dendritic cells from wild-type littermates (Fig. 7B). The levels of IFN-γ produced by the dendritic cells after IL-23 stimulation were below detection limit (Fig. 7B). The LPS-stimulated dendritic cells from miR-21–deficient mice produced significantly increased levels of IFN-γ compared with dendritic cells from wild-type littermates (Fig. 7B), likely because of the increased IL-12p70 production from miR-21–deficient dendritic cells after LPS stimulation (Fig. 5A).

We aimed to broaden our findings by testing whether miR-21 had a role in regulating IL-12/IFN-γ responses in another tissue outside the lung and also in a distinct, nonallergic model, characterized by polarized Th1 responses. Accordingly, we used an NP-OSu–induced Th1cytokine-associated model of delayed-type hypersensitivity in the skin (28). In this model, sensitized mice were challenged by footpad injection of NP-OSu. Compared to wild-type littermate controls, miR-21^−/− mice had an increase in footpad thickness and footpad weight, indicative of an increased footpad swelling response in these mice (Fig. 8A, 8B). This was accompanied by an increase in IFN-γ mRNA in the footpad, with a corresponding increase in the IFN-γ–induced chemokine CXCL10 (Fig. 8C, 8D) (29).

In this study, we provide definitive evidence that the main function of miR-21 is to regulate the IL-12/IFN-γ axis in multiple models of immune hypersensitivity and that the role of miR-21 in this process is nonredundant. Notably, we provide empiric evidence that this is a major and functionally relevant role for miR-21. To the best of our knowledge, this is the first study demonstrating the role of miR-21 in regulating Th1 versus Th2 responses in vivo. Prior studies concerning miR-21 have primarily focused on its role in tumorigenesis and tissue remodeling (21, 30–33), so the finding that miR-21 dominantly regulates the IL-12/IFN-γ axis in vivo is unexpected. Prior work concerning the regulation of IL-12/IFN-γ has all focused on their molecular regulation by cytokines and transcription factors (7, 34), so our finding that this critical step in adaptive immunity is regulated by a specific miR presents a new paradigm with broad and deep implications.

We have previously shown that miR-21 is the most upregulated miR in multiple models of experimental allergic airway inflammation (6), a finding verified by a recent report from the laboratory of Dr. Paul Foster (5). We previously proposed that miR-21 targets IL-12p35 based on predictive algorithms and in vitro studies using reporter systems in heterologous cells. However, the in vivo function of miR-21 in the setting of natural adaptive immune responses had not been established, especially in light of the myriad of miRs dysregulated during allergic airway inflammation. It is also well known that miR-predictive algorithms have limitations, including lack of specificity, as single miRs have the capacity to directly target a myriad of mRNAs. This is illustrated in part by the recent report that miR-21–deficient mice have no alteration in cardiac remodeling, even though this outcome was expected based on predictive algorithms and antagomir studies (21, 30).

IL-12 is a major cytokine that regulates Th1 versus Th2 decisions primarily by inducing T cells to produce the Th1 cytokine IFN-γ (7). Using gene expression microarray and systems biology analysis, we identified an IL-12/IFN-γ–dependent pathway as the most prominent upregulated pathway in the lungs of OVA-challenged miR-21^−/− mice compared with wild-type littermate controls, providing substantial evidence that this is the major pathway dysregulated in the miR-21–deficient mice. We then demonstrated that the miR-21–deficient mice have increased levels of the Th1 cytokine IFN-γ and relevant IFN-γ–induced genes, such as CXCL9, after allergen challenge compared with wild-type littermates. CD4⁺ T cells from OVA-challenged miR-21^−/− mice produced increased levels of IFN-γ and reduced levels of IL-4 compared with cells from wild-type littermates following OVA or anti-CD3/CD28 stimulation, indicating the functional consequences of IL-12 production on subsequent T cell responses. By contrast, naive CD4⁺ T cells from miR-21–deficient mice produced similar levels of IFN-γ and IL-4 compared with cells from wild-type littermates following anti-CD3/CD28 stimulation, indicating that the Th1 skewed response in miR-21^−/− mice is likely due to increased IL-12 production. Indeed, bone marrow-derived dendritic cells from miR-21–deficient mice produced significantly more IL-12p70 compared with control mice, thus providing a mechanistic basis for the experimental asthma phenotype seen in the miR-21^−/− mice. Because NK cells and dendritic cells are also important IFN-γ–producing cells, we measured the levels of IFN-γ production by NK cells and dendritic cells from miR-21–deficient mice compared with wild-type littermates. The NK cells and dendritic cells from miR-21–deficient mice produced similar amounts of IFN-γ compared with wild-type littermates following IL-12 stimulation, further supporting our hypothesis that the Th1-skewed response in miR-21^−/− mice is due to increased IL-12 production. The LPS-stimulated dendritic cells from miR-21–deficient mice produced significantly increased levels of IFN-γ compared with dendritic cells from wild-type littermates, likely because of the increased IL-12p70 production from miR-21–deficient dendritic cells after LPS stimulation.

We evaluated the applicability of these findings in an independent system in vivo by examining the Th1-dependent, delayed-type hypersensitivity cutaneous response in miR-21–deficient mice. Interestingly, miR-21–deficient mice had an enhanced delayed-type hypersensitivity response characterized by increased footpad swelling and IFN-γ level in the footpad. Thus, we propose that miR-21–mediated attenuation of the IL-12/IFN-γ pathway has profound effects on regulating adaptive immune responses. These findings underscore the sensitivity of Th1 versus Th2 decisions and identify a key role for miR-mediated posttranscriptional regulation of this process.

Our findings do not preclude the possibility that miR-21 may regulate multiple other pathways. Abundant evidence points to the role of miR-21 in regulating antiapoptosis pathways, particularly those involved in oncogenesis, by targeting genes such as programmed cell death 4 and phosphatase and tensin homology deleted from chromosome 10, which have also been implicated in immunity (35–38). To consider the potential significance of miR-21 in the polarization of the immune-inflammatory responses in normal and cancerous tissues, we evaluated miR-21 expression in two large human Affymetrix gene expression datasets (Affymetrix), one based on a series of >2000 human cancers (the Cancer ExPO database National Center for Biotechnology Information GSE2109, compiled by the international genomics consortium) and the other based on the Human Body Index, GSE7307, that contains 677 samples representing >90 distinct human tissue types. We identified the Affymetrix probe set 224917_at as the best potential indicator of miR-21 gene expression and used Pearson correlation analysis across the cancer set to identify 269 probesets corresponding to 182 human genes for which the pattern of expression was highly similar to that of the miR-21 host gene (Supplemental Table IV). Of these, 245 probe sets maintained a correlation level >0.2 in the normal human body map. Tissues that exhibited a high level of miR-21 expression included many mucosal epithelium-containing tissues such as lung and oral mucosa as well as the lymphoid tissues spleen, tonsil, and lymph nodes. Remarkable similarities were also identified between the human signature and mouse pathways, with both being enriched in immunoinflammatory processes (Table I). It is notable that a large-scale survey to determine the miRNA signature of 540 diverse tumor samples identified miR-21 as the only miRNA upregulated in all of these tumors (9); and miR-21 has been recently shown to be a true oncomiR (31, 32). Given that Th2 responses have recently been shown to be tumorigenic (39, 40), the ability of miR-21 to promote a Th2 versus Th1 microenvironment may be an additional mechanism for its oncogenic effects.

In summary, we have identified miR-21 as a regulator of the polarization of Th1/Th2 responses in multiple models of immunoinflammatory hypersensitivity. We propose that despite the multistep nature of adaptive immune responses and the number of regulatory steps involved, Th1 versus Th2 decisions appear to be miR-21 dependent. The presented findings have broad implications for diseases associated with imbalanced Th1 and Th2 responses (e.g., asthma) and other diseases associated with miR-21 overexpression (e.g., cancer). Thus, targeting miR-21 and understanding variations in its activity and impact on the IL-12/IFN-γ pathway may lead to potentially far-reaching treatments and preventions. Taken together, our findings provide a new mechanism for regulating a critical arm of adaptive polarized immune responses.

We thank Shawna K.B. Hottinger for editorial assistance.

M.E.R. has an equity interest in reslizumab, a drug being developed by Cephalon, Inc. The other authors have no financial conflicts of interest.