miR-708 Negatively Regulates TNFα/IL-1β Signaling by Suppressing NF-κB and Arachidonic Acid Pathways

Two pathways commonly dysregulated in autoimmune diseases and cancer are tumor necrosis factor alpha (TNFα) and interleukin 1 beta (IL-1β) signaling. Researchers have also shown that both signaling cascades positively regulate arachidonic acid (AA) signaling. More specifically, TNFα/IL-1β promotes expression of the prostaglandin E2- (PGE2-) producing enzymes, cyclooxygenase-2 (COX-2) and microsomal prostaglandin E synthase-1 (mPGES-1). Exacerbated TNFα, IL-1β, and AA signaling have been associated with many diseases. While some TNFα therapies have significantly improved patients' lives, there is still an urgent need to develop novel therapeutics that more comprehensively treat inflammatory-related diseases. Recently, researchers have begun to use RNA interference (RNAi) to treat various diseases in the clinic. One type of RNAi is microRNA (miRNA), a class of small noncoding RNA found within cells. One miRNA in particular, miR-708, has been shown to target COX-2 and mPGES-1. Previous studies have also suggested that miR-708 may be a negative regulator of TNFα/IL-1β signaling. Therefore, we studied the relationship between miR-708, TNFα/IL-1β, and AA signaling in diseased lung cells. We found that miR-708 negatively regulates TNFα/IL-1β signaling in nondiseased lung cells, which is lost in diseased lung cells. Transient transfection of miR-708 suppressed TNFα/IL-1β-induced changes in COX-2, mPGES-1, and PGE2 levels. Moreover, miR-708 also suppressed TNFα/IL-1β-induced IL-6 independent of AA signaling. Mechanistically, we determined that miR-708 suppressed IL-6 signaling by reducing expression of the nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) activator inhibitor of nuclear factor kappa-B kinase subunit beta (IKKβ). Collectively, our data suggest miR-708 regulates TNFα/IL-1β signaling by inhibiting multiple points of the signaling cascade.

Loss of TNFα signaling control has been shown to contribute to numerous diseases, including rheumatoid arthritis (RA), Crohn's disease, atherosclerosis, cancer, and other autoimmune diseases [3]. Similarly, dysregulated IL-1β expression contributes to autoimmune diseases and cancer [5]. There are several approved TNFα inhibitors and one IL-1R antagonist used to treat autoimmune diseases, highlighting the importance of this pathway in chronic inflammatory diseases. While these inhibitors have improved patient outcomes, 20-40% of RA patients are nonresponsive to TNFα inhibitors and resistance amongst responders is increasing [6]. More recently, researchers have begun to study the therapeutic potential of TNFα and IL-1β inhibition in cancer. Studies have shown that TNFα/IL-1β inhibition can suppress protumorigenic immune signaling through reduced angiogenesis, metastasis, and immune evasion [7,8]. While previous work has highlighted the importance of these cytokines in disease, more efficacious therapies are still needed.
It was recently shown that cyclooxygenase-2 (COX-2), the rate-limiting enzyme in the arachidonic acid signaling pathway, was constitutively overexpressed in ulcerative colitis patients who did not respond to TNFα inhibitors [9]. Moreover, COX-2 is also overexpressed in RA and atherosclerosis [10]. Overexpression of COX-2 leads to exacerbated prostaglandin E2 (PGE 2 ) production. While PGE 2 is important for hematopoietic stem cell regeneration, inflammation, and gut integrity, increased levels can promote proliferation, invasion, survival, angiogenesis, and immune evasion in cancer [11,12]. It has also been well documented that TNFα and IL-1β signaling promotes COX-2 expression [13][14][15]. Lastly, there is also significant crosstalk between TNFα, IL-1β, COX-2, and IL-6, another proinflammatory cytokine implicated in disease [16]. Given these data, it would be worthwhile to examine whether novel COX-2-inhibiting therapies can more efficaciously treat TNFα/IL-1β-related diseases.
One novel method to inhibit gene expression is through microRNA (miRNA). miRNAs are small (21-25 nts), noncoding RNAs that regulate gene expression posttranscriptionally [17]. miRNAs generally suppress gene expression by targeting the transcript's 3 ′ UTR with incomplete complimentary, resulting in transcript degradation or translational stalling [17]. One miRNA in particular, miR-708-5p (miR-708), has been shown to target COX-2 in diseased lung cells [18]. miR-708 was also shown to target the NF-κB activator inhibitor of nuclear factor kappa-B kinase subunit beta (IKKβ), which is intricately involved in canonical TNFα signaling [19]. Additionally, forced overexpression of miR-708 in human airway smooth muscle cells decreased expression of asthma-related genes [20]. miR-708 also decreased alcohol-induced liver inflammation by targeting zinc finger E-box-binding homeobox 1 (ZEB-1) upregulation of TNFα and IL-6 [21]. Lastly, injections of miR-708 mimic into a rodent RA model ameliorated the RA index by inhibiting WNT signaling [22]. Given the various immunoregulatory functions of TNFα, IL-1β, IL-6, and COX-2, paired with the anti-inflammatory functions of miR-708, we decided to investigate whether miR-708 negatively regulates TNFα and IL-1β signaling in disease.
In this article, we show that miR-708 expression is temporally induced by TNFα/IL-1β in nondiseased lung cells. Conversely, miR-708 expression is nonresponsive to TNFα/IL-1β in diseased lung cells. Exogenous miR-708 in nonresponsive lung cells reduced TNFα/IL-1β-induced changes in COX-2, mPGES-1, PGE 2 , and IL-6 levels. Lastly, we determined that miR-708 is inhibiting TNFα/IL-1β sig-naling dually by suppression of AA and NF-κB signaling in diseased lung cells. Collectively, these data suggest miR-708 is a potent negative regulator of TNFα/IL-1β, which is lost in diseased lung cells.
2.2. Enzyme-Linked Immunosorbent Assay (ELISA). A549 PGE 2 levels in cell culture media were analyzed using the PGE 2 Express ELISA Kit (500141, Cayman Chemical, Ann Arbor, MI) per manufacturer's instructions. Media was removed, and cells were incubated for 20 min with serumfree media containing 10 μM arachidonic acid (Cayman Chemical) in serum-free DMEM. Collected media was centrifuged at 5000 g × 10 min, 4°C. Media was transferred to new tubes and then centrifuged at 2000 g × 10 min, 4°C, before being transferred to new tubes. Before analysis, samples were diluted 10x with 1x ELISA buffer. Absorbance was read using the SpectraMax M2 plate reader (Molecular Devices, San Jose, CA). PGE 2 levels were measured in technical duplicates, normalized to total protein levels, and are an average of ≥3 biological replicates. When probing for IL-6, we used a similar kit from Cayman Chemical (501030) in accordance with the manufacturer's protocol. In contrast to the PGE 2 ELISA, complete media was collected from cells at each time point indicated. Centrifugation was performed as stated above. Media was not diluted prior to performing the IL-6 ELISA. IL-6 levels were measured in technical duplicates, normalized to total protein levels, and are an average of ≥3 biological replicates.
3.2. miR-708 Expression Is Nonresponsive to TNFα/IL-1β in Diseased Lung Cells. We repeated our experiments in A549 (diseased) lung cells, which are a lung adenocarcinoma cell line. While cancerous, researchers have used this cell line to study other inflammatory-related diseases. A549 cells treated with TNFα/IL-1β dramatically increased COX-2 protein levels through 48 hours, with no resolution observed ( Figure 2(a)). Conversely, we saw nonsignificant TNFα/IL-1β-induced increases in miR-708 expression in A549 cells (Figure 2(b) p = n:s:, n = 3). When paired with Beas2b miR-708 expression changes, we see that A549 miR-708 levels are significantly lower at baseline (Figure 2(b), p < 0:0001, n = 3). We have previously shown that reduced miR-708 expression in A549 cells is due to ODZ4 promoter methylation [18]. Moreover, A549 miR-708 levels never increased above 15% of baseline Beas2b expression after TNFα/IL-1β treatment, and miR-708 levels are significantly lower at every time point in A549 cells (Figure 2(b), p < 0:001, n = 3). These data suggest that ODZ4 promoter methylation prevents TNFα/IL-1β-induced miR-708 expression, disconnecting the   Mediators of Inflammation negative feedback loop in diseased lung cells and leading to unresolved COX-2 signaling. As a control, we performed viability (WST-1) analysis on Beas2b and A549 cells untreated or TNFα/IL-1β treated and found that TNFα/IL-1β treatment did not affect cellular viability in either cell line (Supplemental Figure 1). Given these data, we next investigated if transient transfection of exogenous miR-708 could suppress TNFα/IL-1β-induced changes in A549 cells.
3.3. miR-708 Reduces TNFα/IL-1β-Induced Changes in Diseased Lung Cells by Suppressing NF-κB Signaling. We transiently transfected A549 cells with mock (-miR-708) or synthetic miR-708 and then exposed cells to TNFα/IL-1β for 0-48 hours. As seen in Figure 3(a), A549 cells without miR-708 treatment had robust induction of COX-2 protein expression in a time-dependent manner. A549 cells treated with synthetic miR-708 had a significant reduction in both COX-2 and mPGES-1 protein levels across all TNFα/IL-1β time points (Figure 3(a)). We also measured PGE 2 release by ELISA and observed significantly decreased PGE 2 levels at every time point in miR-708-treated A549 cells compared to mock samples ( Figure 3(b), p < 0:05, n ≥ 3). PGE 2 has been shown to form a positive feedback loop to enhance COX-2 expression and other cytokines. We inhibited COX-2 production of PGE 2 with celecoxib (CEL) and repeated our experiments. While CEL efficiently suppressed PGE 2 production (Figure 3(d), p < 0:05, n ≥ 3), it only partially reduced TNFα/IL-1β regulation of COX-2 protein expression (Figure 3(c)). Next, we measured IL-6 production, as researchers have shown that it is directly regulated by COX-2/mPGES-1-derived PGE 2 . While miR-708 prevented TNFα/IL-1β-induced IL-6 expression, CEL had no effect on IL-6 levels after TNFα/IL-1β treatments ( Figure 4, p < 0:05, p = n:s:, n ≥ 3). These data reveal that miR-708 is suppressing IL-6 production independent of AA signaling. As previously discussed, TNFα/IL-1β has been shown to induce IL-6 both through AA signaling and independent of COX-2, mPGES-1, or PGE 2 . Given these data, we next examined the mechanism by which miR-708 is suppressing IL-6 levels in diseased lung cells.
To achieve this, we utilized the pNL3.2.NF-κB-RE reporter luciferase assay. Briefly, this reporter construct has an NF-κB response element (RE) upstream of the NLucP luciferase gene. Once NF-κB is shuttled to the nucleus, it binds to the NF-κB RE, activating NLucP expression. Changes in NLucP expression can then be quantified via luminescence assays. We transfected A549 cells with mock, 25 nM miR-708, or 10 μM BAY-11-7082, followed by the pNL3.2.NF-κB-RE construct the following day. The next morning, we treated these cells with TNFα/IL-1β for 0 or 5 h and then measured luminescence. We found that TNFα/IL-1β+mock A549 cells had a 16x increase in luminescence after 5 hours (Figure 6(a)). While BAY-11-7082 significantly reduced TNFα/IL-1β-mediated NF-κB activation and luminescence by 28%, miR-708 had a greater suppressive effect than BAY-11-7082 ( Figure 6(a), p < 0:0001, n = 3). These data suggest miR-708 is indeed suppressing TNFα/ IL-1β-mediated NF-κB activation and downstream signaling. Given these data, we repeated the PGE 2 and IL-6 ELISAs in A549 cells. We found that BAY-11-7082 significantly reduced A549 PGE 2 production after TNFα/IL-1β treatment similarly to miR-708 ( Figure 6(b), p < 0:0001, n = 3). Lastly, BAY-11-7082 also suppressed IL-6 production in A549 cells post-TNFα/IL-1β treatment in a time-dependent manner ( Figure 6(c), p < 0:01, n = 3). There was not a significant difference in IL-6 expression between miR-708 and BAY-11-7082 treatments at every time point, suggesting that miR-708's effects can be attributed to NF-κB signaling inhibition. Collectively, these data reveal that miR-708 is suppressing IL-6 production independent of AA signaling in lung cancer cells. IL-6 is not a predicted miR-708 target, suggesting miR-708's regulation of IL-6 is indirect. In Figure 7, we propose a model for how miR-708 is suppressing TNFα/IL-1β signaling in A549 cells based on our data and published studies. We surmise that miR-708 suppresses TNFα/IL-1β through multiple mechanisms, by inhibiting both NF-κB and AA signaling in diseased lung cells.

Mediators of Inflammation
In this study, we first confirmed a previous work that showed TNFα upregulated COX-2 and mPGES-1 protein expression within 6 hours and returned to baseline by 48 hours in normal lung cells (Figure 1(a)). Interestingly, miR-708 was temporally upregulated similar to other TNFαdependent genes in nondiseased lung cells (Figure 1(b)). In diseased lung cells, TNFα/IL-1β highly upregulated COX-2 expression, but AA signaling was never restored to baseline (Figure 2(a)). Conversely, TNFα/IL-1β had no effect on miR-708 expression in lung cancer cells (Figure 2(b)). This suggests that a novel miR-708 negative feedback loop may be lost in diseased lung cells. Therefore, we tested the ability of exogenous miR-708 to suppress TNFα/IL-1β signaling in diseased lung cells.
We found that miR-708 decreased IKKβ protein expression, a potent NF-κB activator ( Figure 5). Using a NF-κB promoter luciferase construct, we determined that miR-708 significantly reduced TNFα/IL-1β-induced NF-κB transcriptional activity (Figure 6(a)). Lastly, we inhibited NF-κB signaling using BAY-11-7082 and found that NF-κB inhibition reduced PGE 2 and IL-6 production similarly to miR-708 treatment (Figures 6(b) and 6(c)). This finding supports the hypothesis that miR-708's regulation of IL-6 is through the NF-κB pathway, not AA signaling. Given these data, we created a model showing miR-708's dampening effects on TNFα/IL-1β in lung cells (Figure 7). We found that miR-708 acts dually to suppress TNFα/IL-1β signaling by inhibiting both AA and NF-κB signaling. Moreover, loss of miR-708 in diseased lung cancer may contribute to uncoupled TNFα/IL-1β signaling. Lastly, these data suggest miR-708 may act as a novel therapeutic to treat TNFα/IL-1β-related diseases. These data provide the foundation for further exploring the role of miR-708 in TNFα/IL-1β-related diseases.
While we have laid the foundation for expanding research on the precise functions of miR-708 in TNFα/IL-1β signaling, several questions remain. First, how is TNFα inducing miR-708 expression in normal lung cells? Previous researchers have shown that encourages the expression/activity of the miR-708 regulators CHOP, CTBP2, C/EBP-β, and CTCF [31][32][33][34]. Given these data, there are multiple avenues by which TNFα/IL-1β may stimulate miR-708 expression. Given that TNFα suppresses GRα, MYC, and E2F1 signaling, known miR-708 regulators, it is unlikely that these transcription factors are involved in our study. On the other hand, CHOP and C/EBP-β stimulate miR-708 expression and are TNFα-responsive transcription factors. Therefore, CHOP and C/EBP-β are possibly transcriptional regulators of TNFα/IL-1β-induced miR-708 expression. While this provides a plausible signaling axis, we have not identified why miR-708 is not upregulated after TNFα/IL-1β treatment in diseased lung cells.
There are several possible explanations for how TNFα/IL-1β-mediated activation of miR-708 is lost in diseased lung cells. First, transcriptional activators may not be expressed in these cells. Second, the ODZ4 promoter may be inaccessible to transcription factors due to promoter methylation and chromatin remodeling. Third, miR-708negative regulators may be preventing transcriptional activation. Previous studies have shown that CHOP signaling is inhibited or defective in cancer [35]. Additionally, C/EBP-β is necessary for effective lung cell inflammatory responses and is lost in cancer [36,37]. Therefore, loss of CHOP and C/EBP-β activity may prevent activation of miR-708 expression in diseased lung cells. Secondly, it has been previously shown that the ODZ4 promoter, which miR-708 is under the control of, is hypermethylated in diseased lung cells [18]. This may be through EZH2, which is positively influenced by TNFα/IL-1β signaling and has been shown to hypermethylate the ODZ4 promoter in cancer, resulting in repressed miR-708 expression [38,39]. Lastly, two negative regulators of miR-708 that are positively regulated by TNFα/IL-1β, CTBP2 and CTCF, are overexpressed in lung immune-related diseases [40,41]. Collectively, these studies suggest that the loss of miR-708 responsiveness to TNFα/IL-1β stimulation may be a multifaceted mechanism. Loss of transcriptional activators, overexpression of negative regulators, and ODZ4 promoter hypermethylation may collectively inhibit miR-708 expression in diseased lung cells. The consequences of this miR-708 suppressive network 7 Mediators of Inflammation may be profound. TNFα-induced chronic inflammatory states frequently result in epithelial cell transformation. Loss of TNFα-negative regulators such as miR-708 further propagates TNFα signaling, increasing the risk of transformation and autoimmune disease. Therefore, miR-708 may be a crucial TNFα/IL-1β signaling-suppressing miRNA involved in early stage inflammatory-related oncogenesis and autoimmunity. Given these proposed circumstances, it would be justifiable to study the therapeutic potential of miR-708 in inflammatory-related cancers and autoimmune diseases.
Given the data we have uncovered, as well as TNFα/IL-1β's therapeutic relevance in autoimmune diseases and cancer, it would be beneficial to examine miR-708's ability to reduce autoimmune-related inflammation and oncogenesis. TNFα, IL-1β, and IL-6 are all prominent genes implicated in immune evasion. The ability of miR-708 to suppress TNFα, IL-1β, IL-6, and AA signaling axes may provide a more comprehensive therapeutic intervention compared to current targeted therapies. Additionally, significant subsets of patients receiving TNFα inhibitors become resistant to treatment. In these populations, researchers showed that COX-2 was overexpressed [9]. Moreover, COX-2 inhibition overcame TNFα resistance in these ulcerative colitis patients [9]. While COX-2 inhibitors have shown promise in treating autoimmune diseases and cancer, they have unacceptable long-term side effect profiles. Therefore, using miR-708 to treat these populations may provide a fresh therapeutic alternative. While the efficacy and safety of miR-708 is unknown, future studies will uncover the therapeutic potential of miR-708 in treating TNFα/IL-1β-related diseases.  (c) Figure 6: miR-708 suppresses PGE 2 and IL-6 secretion by inhibiting NF-κB signaling in diseased lung cells. (a) Relative NF-κB promoter reporter NanoLuc luciferase activity in A549 cells. Relative luciferase activities were measured in response to mock (blue), 25 nM miR-708 (red), or 10 μM BAY-11-7082 (green) treatment for 0 or 5 h. Data were normalized to total protein and mock 0 h samples. Differences were compared between samples within 0 and 5 h time points. (b) ELISA measuring exogenous PGE 2 levels in A549 cells cotreated with vehicle (blue), 25 nM miR-708 (red), or 10 μM BAY-11-7082 (green) and TNFα/IL-1β for 0-48 hours. (c) ELISA measuring exogenous IL-6 levels in A549 cells cotreated with vehicle (blue), 25 nM miR-708 (red), or 10 μM BAY-11-7082 (green) and TNFα/IL-1β for 0-48 hours. * * p < 0:01 and * * * * p < 0:0001, n ≥ 3.

Conclusions
Our data suggests miR-708 is a potent negative regulator of TNFα/IL-1β signaling in lung epithelial cells. Loss of endogenous miR-708 expression may contribute to unchecked TNFα/IL-1β signaling, leading to inflammatory-related diseases. Reestablishing miR-708 levels may counteract TNFα/IL-1β-induced inflammation, ultimately resolving disease phenotype.

Conflicts of Interest
The authors declare that they have no conflicts of interest.