Interaction of RNA-binding protein HuR and miR-466i regulates GM-CSF expression

Granulocyte-macrophage colony-stimulating factor (GM-CSF) produced by T helper 17 (Th17) cells plays an essential role in autoimmune diseases. Transcriptional regulation of Th17 cell differentiation has been extensively studied, but post-transcriptional regulation of Th17 cell differentiation has remained less well characterized. The RNA-binding protein HuR functions to promote the stability of target mRNAs via binding the AU-rich elements of the 3′ untranslated region (3′UTR) of numerous pro-inflammatory cytokines including IL-4, IL-13, IL-17 and TNF-α. However, whether HuR regulates GM-CSF expression in Th17 cells has not been fully investigated. Here we showed that HuR conditional knockout (KO) Th17 cells have decreased GM-CSF mRNA in comparison with wild-type (WT) Th17 cells, and that HuR binds directly to GM-CSF mRNA 3′UTR. Interestingly, HuR deficiency increased the levels of certain microRNA expression in Th17 cells; for example, miR-466i functioned to mediate GM-CSF and IL-17 mRNA decay, which was confirmed by in vitro luciferase assay. Furthermore, we found that HuR promoted Mxi1 expression to inhibit certain miRNA expression. Taken together, these findings indicate that interaction of HuR and miR-466i orchestrates GM-CSF expression in Th17 cells.

It is well known that interaction of HuR and microRNAs (miRNAs) controls target mRNA expression in response to environmental stimuli 31,32 . MiRNAs are small, non-coding RNAs with approximately 21 to 24 nucleotides (nt) 33 , which regulate the expression of numerous target genes by mediating their mRNA decay and/or repressing their translation [33][34][35] . The most common motif is perfect pairing between nucleotides 2 and 7 at the 5′ end of the miRNA, which is called the 'seed' region. It is thought that most miRNA-mRNA interactions involve the seed region at the 5′ end of miRNA. The small size of miRNAs provides a limited amount of information for specificity. Furthermore, as partial pairing between a miRNA and a target site is often sufficient for miRNA function, which means that a single miRNA can regulate multiple mRNAs but which also makes target predictions complicated 36 . Although cells contain hundreds of miRNAs, only a limited number of miRNAs have been validated functionally. Given the importance of miRNAs in regulating expression of numerous genes, further characterizing specific miRNA function will improve our understanding of gene regulation. Here we provide evidence that HuR post-transcriptionally modulates GM-CSF expression in Th17 cells. The level of miR-466i is increased in HuR KO Th17 cells compared with WT Th17 cells. The 3′UTRs of GM-CSF and IL-17 mRNA are potential targets of miR-466i, as shown by Targetscan analysis. In vitro luciferase transfection assay demonstrated that miR-466i could target GM-CSF and IL-17 mRNA 3′UTRs for decay, suggesting that miR-466i has potential as a novel reagent for therapeutic intervention in autoimmune inflammation.

Knockout of HuR reduces GM-CSF expression in Th17 cells.
Our previous studies showed that the levels of HuR mRNA and protein correlate with IL-17 mRNA and protein expression in Th17 cells and that knockout of HuR reduces IL-17 mRNA and protein levels in Th17 cells 29 . Considering that GM-CSF plays a non-redundant role in Th17 cell induction of autoimmune demyelination 7,14 , we were interested in determining if HuR regulates GM-CSF mRNA expression in Th17 cells. To address this question, we used HuR conditional KO (OX40-Cre HuR f/f ) mice 29 . Naive CD4 + T cells do not express OX40, but activated T cells do 36 , thus deleting HuR. An advantage of using this approach is that conditional knockout of HuR does not affect early development of T cells. Naive CD4 + T cells were isolated from wild type (WT) and KO mouse splenocytes, then were cultured under Th0 and Th17 cells-polarizing conditions. Western blot assays were performed to quantify HuR levels. Endogenous HuR was gradually degraded after cell activation in KO CD4 + T cells from day 3 to 5 of culture, and HuR was barely detected at day 5 29 . Quantitative real-time PCR (qRT-PCR) data showed that KO Th0 and Th17 cells had significantly less GM-CSF mRNA as compared with WT counterparts on day 4 of culture ( Fig. 1a,b). Flow cytometry analysis also showed that the GM-CSF protein was also reduced in HuR KO Th0 and Th17 cells in comparison with WT cells (Fig. 1c). Results were confirmed by ELISA assay (Fig. 1d,e). These results demonstrated that knockout of HuR reduced GM-CSF mRNA and protein in Th0 and Th17 cells.
HuR binds to GM-CSF mRNA 3′UTR. The regulation of GM-CSF mRNA expression by HuR suggested that GM-CSF mRNA might be a direct target of HuR. To study the function of HuR in regulation of GM-CSF in Th17 cells, we did RNA immunoprecipitation (RIP) assay to test if HuR protein binds to GM-CSF mRNA. The association of GM-CSF mRNA with HuR was tested by isolating RNA from the ribonucleoprotein (RNP) complexes with anti-HuR Ab (Fig. 2a). Our results showed a remarkable GM-CSF mRNA enrichment in the anti-HuR immunoprecipitation (IP) sample compared with an isotype-matched IgG control (Fig. 2b). IL-23R mRNA was not enriched in the HuR pulldown sample as a negative control (Fig. 2b). To confirm HuR interaction with GM-CSF mRNA, we utilized biotin pulldown, a second independent method. Two biotinylated transcripts that are complementary to sequences in GM-CSF mRNA ORF (open reading frame) and 3′ UTR were generated according to a protocol described in Materials and Methods (Fig. 2c). The biotinylated transcripts were incubated with cytoplasmic protein lysates from NIH/3T3 cells, which contain large amounts of HuR protein. The result of western blot showed that HuR was clearly detected in the samples precipitated by the biotinylated transcript of GM-CSF mRNA 3′ UTR ( Fig. 2c-e). HuR protein could not be detected in the GAPDH 3′ UTR biotinylated probe pulldown complex (Fig. 2e), and this transcript has no putative HuR binding sites 29 . Taken together, these results indicated that HuR directly binds to 3′ UTR of GM-CSF mRNA, suggesting a regulatory relationship. Thus, HuR binds to and stabilizes GM-CSF mRNA, resulting in increased mRNA accumulation in Th17 cells.

HuR inhibits miR-466i expression in Th17 cells.
Since miRNAs play an important role in mediating target mRNA decay, decreasing mRNA levels by miRNAs represents a better approach to disrupt their protein function. We used several miRNA analysis programs to explore miRNAs that have the potential to target GM-CSF and IL-17 mRNAs. Interestingly, by Targetscan analysis, we found many miRNAs that could potentially bind to GM-CSF and IL-17 mRNAs, respectively (Supplementary Table 1). Among these miRNAs, it is only miR-466i that has potential binding sites in both GM-CSF and IL-17 mRNAs (Fig. 3a-d). Previous work showed that overexpression of miR-446i upregulates IL-10 expression in macrophages by antagonizing RNA-binding protein tristetraprolin-mediated IL-10 mRNA degradation, and miR-466i inhibits antiviral innate immune response by targeting interferon-α 37,38 . A recent study showed that miR-466i plays a bipolar role in inflammation, promoting an acute inflammatory response initially and enhancing resolution during the late stage of inflammation 39 . However, whether miR-466i mediates GM-CSF and IL-17 mRNA degradation remains unknown, thus, making this an important question to address.
Interaction of RNA-binding protein HuR with miRNAs post-transcriptionally regulates gene expression 32 . Our previous studies showed that HuR post-transcriptionally modulates IL-17 mRNA expression and promotes autoimmune neuroinflammation 29 . Knockout of HuR reduces IL-17 mRNA half-life and destabilizes it. Here, we found that HuR directly regulated GM-CSF production by binding to 3′UTR of GM-CSF mRNA ( Fig. 2b and e). We therefore performed qRT-PCR to examine miR-466i expression in WT and HuR KO Th17 cells. Interestingly, miR-466i was highly expressed in HuR KO Th17 cells in comparison with WT Th17 cells, but there was no significant change in miR-155 expression (Fig. 4a), suggesting that miR-466i may play an active role in mediating GM-CSF and IL-17 mRNA decay in the absence of HuR, and that HuR inhibits expression of certain miRNAs 27 .
To understand how HuR negatively regulates expression of certain miRNAs, we checked the expression of some molecules related to miRNA expression in HuR KO and WT Th17 cells. It is well known that Drosha, Dicer, Ago1/2 and c-Myc are actively involved in miRNA expression 40 , but their expression showed no significant change in the presence or absence of HuR (data not shown). However, expression of Mxi1 mRNA and protein, a repressor of c-Myc 41 , was significantly reduced in the absence of HuR compared with WT Th17 cells ( Fig. 4b-d).
Thus, it is possible that HuR promotes Mxi1 expression, which suppresses c-Myc function, resulting in its indirectly inhibiting expression of certain miRNAs. In line with this notion, knockdown of Mxi1 in WT Th17 cells by Mxi1 siRNA transfection increased miR-466i expression but decreased GM-CF mRNA compared with counterparts treated with scramble siRNA (Fig. 4e,f).

MiR-466i-mediates GM-CSF and IL-17 mRNA decay by interacting with their 3′UTRs.
To further investigate whether miR-466i modulates GM-CSF and IL-17 expression by degrading it through interacting with their 3′UTRs, a reporter vector with firefly and Renilla Dual-Luciferase (RDL) containing GM-CSF 3′UTR Figure 1. Knockout of HuR reduces GM-CSF mRNA in CD4 + T cells. Naive WT and HuR KO CD4 + T cells were isolated as described in Materials and Methods and cultured with anti-CD3 and anti-CD28 stimulation in Th0 and Th17 cell polarizing conditions. Expression of GM-CSF mRNA (a,b) and protein (c-e) in Th0 and Th17 cells was measured by qRT-PCR, flow cytometry, and ELISA, respectively. Knockout of HuR reduced the GM-CSF mRNA and protein levels in Th0 and Th17 cells compared with WT counterparts. Results in (a,b,d and e) represent the summary of three independent experiments. The representative of flow analysis was shown in (c). **p < 0.01, ***p < 0.001. was co-transfected with miR-466i mimics into HeLa cells (Fig. 5a). Overexpression of miR-466i decreased luciferase activity of the reporter construct by analysis of firefly luciferase activity which was normalized by Renilla luciferase activity, but overexpression of scramble miRNAs did not (Fig. 5a), suggesting that miR-466i functions to degrade GM-CSF mRNA through binding to 3′UTR. Similarly, overexpression of miR-466i in HeLa cells transfected with RDL vector containing IL-17 mRNA 3′UTR also reduced luciferase activity (Fig. 5b). These results thus suggest that miR-466i mediates GM-CSF and IL-17 mRNA decay through binding to their 3′UTRs. In line with these results, overexpression of miR-466i in WT Th17 cells by transfection reduced expression of GM-CSF and IL-17 mRNA compared to that with scramble miRNA (Fig. 5c).

MiR-466i regulates GM-CSF protein expression in macrophages.
To further confirm the result that miR-466i modulates GM-CSF expression, macrophages were transfected with miR-466i and scramble miR-NAs. GM-CSF mRNA and protein levels were examined by qRT-PCR and flow cytometry, respectively. The data showed that expression of GM-CSF mRNA was remarkably decreased when macrophages were transfected with miR-466i compared with scramble miRNA (Fig. 6a); however, the IL-6 mRNA level was moderately increased in macrophages transfected with miR-466i, which is consistent with a previous report 39 . The GM-CSF protein level was also decreased in macrophages transfected with miR-466i mimics compared with scramble miRNA (Fig. 6b,c). This result indicated that miR-466i also regulates GM-CSF expression in macrophages.
Taken together, these results suggest that HuR and miR-466i interaction modulates GM-CSF expression in Th17 cells. MiR-466i targets GM-CSF and IL-17 mRNA decay by binding to their 3′UTRs, a finding that may lead to development of miR-466i as a novel therapeutic intervention for autoimmune inflammation.

Discussion
MiRNAs are small (approximately 22 bases) non-protein-coding RNAs that recognize target sequences of imperfect complementarity in cognate mRNA and either destabilize them or inhibit protein translation. In principle, gene expression is post-transcriptionally regulated under the control of RNA-binding protein and miRNA interaction 32 . Our current study shows that HuR binds to GM-CSF and IL-17 mRNA by inhibiting miR-466i expression and by preventing miR-466i-mediated targeting of GM-CSF and IL-17 mRNA decay.
In this study, we first used OX40-Cre HuR f/f mice to demonstrate that HuR promotes GM-CSF mRNA expression. Ablation of HuR reduced GM-CSF mRNA and protein expression in Th0 and Th17 cells. We previously reported that knockout of HuR impacted IL-17 expression 29 . GM-CSF and IL-17 in Th17 cells are thus similarly post-transcriptionally regulated by HuR. Here we found that HuR functions to bind to GM-CSF mRNA 3′UTR. HuR ablation therefore reduced GM-CSF mRNA expression, which is consistent with previous reports showing that GM-CSF can be controlled by a post-transcriptional mechanism through ARE in its 3′UTR and stabilized by HuR [42][43][44][45] . However, the mechanism by which GM-CSF mRNA was degraded in the absence of HuR had not been completely characterized in previous reports. Thus, our current study extended previous findings by showing that HuR deficiency in T cells using conditional knockout mice deceased GM-CSF production and increased miR-466i expression. In vitro analysis suggested that miR-466i has the capacity to mediate GM-CSF and IL-17 mRNA decay. The fact that knockout of HuR increased miR-466i, miR-409 and miR-335 expression 30 is consistent with a previous report showing that overexpression of HuR reduces miR-16 expression, and that HuR inhibits miR-16 targeting of Cox-2 46 . To understand why knockout of HuR increased certain miRNA expression, we also performed anti-HuR immunoprecipitation (IP) western blots to determine whether HuR and Ago2 are physically associated with each other. Not surprisingly, Ago2 could not be detected in the anti-HuR IP complex (data not shown). It has been reported that c-Myc is a transcription factor with the capacity to promote certain miRNA expression 47,48 , and that HuR recruits miRNA let-7 to mediate c-Myc RNA decay 49 . Mxi1 is a transcription repressor that negatively regulates c-Myc function 41 . Although knockout of HuR did not impact the expression of c-Myc mRNA (data not shown), Mxi1 mRNA and protein expression was significantly reduced in the absence of HuR (Fig. 4). Therefore, knockout of HuR reduced Mxi1 expression, which may alleviate its repressive effect on the function of c-Myc, leading to an increase in expression of certain miRNAs. Based on this idea, we speculated that knockdown of Mxi1 in WT Th17 cells would result in increased miR-466i expression. Interestingly, we obtained new data which support this notion (Fig. 4). However, how HuR modulates miRNA biogenesis via Mxi1 only and/or other factors need to be further studied.
By in vitro luciferase assay, we demonstrated that overexpression of miR-466i reduced activity of the vector containing GM-CSF or IL-17 mRNA 3′UTR in transfected Hela cells, suggesting that miR-466i has the potential to target GM-CSF and IL-17 mRNA. Indeed, overexpression of miR-466i in WT Th17 cells and macrophages decreased GM-CSF mRNA and its protein production, respectively (Fig. 6). A previous report showed that miR-466i increased anti-inflammatory cytokine IL-10 production by blocking RNA-binding protein tristetraprolin, which destabilized IL-10 mRNA 37 . A recent report demonstrated that miR-466i could target MyD88 mRNA 3′UTR for decay to downregulate IL-12 production and increase IL-10 production 50 . miR-466i might therefore have the therapeutic potential to target GM-CSF-mediated inflammation and to increase IL-10 production for promoting resolution of inflammation.
Since HuR binds to GM-CSF 3′UTR (Fig. 2), and GM-CSF 3′UTR contains nine potential miR-466i binding sites (Supplementary Table 1), it is likely that HuR and miR-466i might competitively bind to the same target 32 . However, we cannot rule out the possibility that with HuR binding, the space-structure of GM-CSF mRNA 3′UTR changes, which prevents miR-466i from binding to it 32 . In any case, HuR inhibition of miR-466i biogenesis also contributes to the effect of HuR on promoting GM-CSF expression.
Overall, RNA-binding protein HuR post-transcriptionally regulates GM-CSF and IL-17 mRNA expression by binding to target mRNAs. HuR promotes Mxi1 expression to inhibit the expression of certain miRNAs. HuR and miR-466i interaction orchestrates GM-CSF and IL-17 mRNA fate. Taking advantage of the many newly developed small molecule inhibitors for HuR in anti-tumor research 51,52 , and based on our current study on HuR regulation of GM-CSF production, it would be worthwhile to test the effects of these newly developed HuR inhibitors on anti-autoimmune inflammation.
All mice used are on the C57BL/6 background and were breed at the animal facility of Arkansas Bioscience Institute at Arkansas State University and/or the Thomas Jefferson University. Animal experiments were approved by the Arkansas State University (ASTATE) and Thomas Jefferson University (TJU) IACUC committee. We followed the regulation of NIH, ASTATE and TJU according to federal and institutional guidelines.

RNA isolation and Quantitative real time-PCR (qRT-PCR). Cultured cells were collected and total
RNA was extracted using TRIzol (Invitrogen, Life Technologies Corp., Thermo Fisher Scientific). Five hundred ng of RNA was reverse transcribed into cDNA using SuperScript III Kit (Life Technologies, Fisher Thermos Scientific Inc.) according to the manufacturer's protocols. The cDNA was subjected to qRT-PCR analysis using the CFX96 Real-Time PCR Detection system (Bio-Rad, Hercules, CA) with SYBR Green reagent Kit (Invitrogen, Life Technologies Corp., Thermo Fisher Scientific) according to the manufacturer's protocols. The levels of test mRNAs were normalized to the levels of Gapdh mRNA for each sample. Forward and reverse primers for specific murine target genes are listed in Table 1. RNA Immunoprecipitation (RIP). RIP was performed according to published protocols 29,54 . Th17 polarized cells at day 4 to 5 culture were lysed using polysome lysis buffer supplemented with RNase inhibitors and protease inhibitors 55 . Lysates were pre-cleared by adding 30 µg of IgG 1 (BD Bioscience) and 50 µl of Protein-A/G Sepharose beads swollen in NT2 buffer with 5% BSA. Beads were coated by adding either IgG1 (BD Biosciences, San Diego, CA) as control or anti-HuR antibody 3A2, and incubated overnight at 4 °C. After extensive washes of pre-coated Protein-A/G sepharose beads, the pre-cleared lysate was added and incubated for 4 h at 4 °C, and then 30 µg of proteinase K was added to digest protein by incubation at 55 °C for 30 min. The extracted RNA was reverse transcribed into cDNA, and qRT-PCR was preformed to detect the presence of specific target mRNAs.
For transfection of siRNA, naïve WT CD4 + T cells were transfected with Mxi1 siRNA and scramble siRNA (Life Technologies Corp., Thermo Fisher Scientific) using Lonza 4D-nucleofector according to the manufacturer's instructions.
Statistics. Bar graphs in figures represent average values ± SEM unless indicated otherwise. Statistical significance between experimental groups was calculated using a two-tailed unpaired Student t test and is shown in the graphs as follows: *p < 0.05, **p < 0.01, ***p < 0.001.