β-Catenin and TCFs/LEF signaling discordantly regulate IL-6 expression in astrocytes

Background The Wnt/β-catenin signaling pathway is a prolific regulator of cell-to-cell communication and gene expression. Canonical Wnt/β-catenin signaling involves partnering of β-catenin with members of the TCF/LEF family of transcription factors (TCF1, TCF3, TCF4, LEF1) to regulate gene expression. IL-6 is a key cytokine involved in inflammation and is particularly a hallmark of inflammation in the brain. Astrocytes, specialized glial cells in the brain, secrete IL-6. How astrocytes regulate IL-6 expression is not entirely clear, although in other cells NFκB and C/EBP pathways play a role. We evaluated here the interface between β-catenin, TCFs/LEF and C/EBP and NF-κB in relation to IL-6 gene regulation in astrocytes. Methods We performed molecular loss and/or gain of function studies of β-catenin, TCF/LEF, NFκB, and C/EBP to assess IL-6 regulation in human astrocytes. Specifically, siRNA mediated target gene knockdown, cDNA over expression of target gene, and pharmacological agents for regulation of target proteins were used. IL-6 levels was evaluated by real time quantitative PCR and ELISA. We also cloned the IL-6 promoter under a firefly luciferase reporter and used bioinformatics, site directed mutagenesis, and chromatin immunoprecipitation to probe the interaction between β-catenin/TCFs/LEFs and IL-6 promoter activity. Results β-catenin binds to TCF/LEF to inhibits IL-6 while TCFs/LEF induce IL-6 transcription through interaction with ATF-2/SMADs. β-catenin independent of TCFs/LEF positively regulates C/EBP and NF-κB, which in turn activate IL-6 expression. The IL-6 promoter has two putative regions for TCFs/LEF binding, a proximal site located at -91 nt and a distal site at -948 nt from the transcription start site, both required for TCF/LEF induction of IL-6 independent of β-catenin. Conclusion IL-6 regulation in human astrocytes engages a discordant interaction between β-catenin and TCF/LEF. These findings are intriguing given that no role for β-catenin nor TCFs/LEF to date is associated with IL-6 regulation and suggest that β-catenin expression in astrocytes is a critical regulator of anti-inflammatory responses and its disruption can potentially mediate persistent neuroinflammation. Video Abstract Graphical abstract

Astrocytes are highly specialized cells in the CNS with well-established immune and non-immune functions including glutamate uptake, maintenance of the blood brain barrier (BBB), and secretion of immune modulators, including IL-6 [20,[39][40][41][42][43][44][45]. We previously reported that astrocytes have robust expression of the Wnt/β-catenin signaling pathway and its disrupting, in response to inflammation and or viral infection, leads to dysregulation of key functions of astrocytes including induction of astrocyte senescence [41,46]. IL-6 is a hallmark cytokine associated with the senescence associated secretory phenotype (SASP), a collection of cytokines and other secretory factors secreted under cellular senescence [47][48][49][50][51]. We therefore assessed the molecular regulation of IL-6 in astrocytes and whether IL-6 is regulated by the Wnt/β-catenin pathway in astrocytes.
β-catenin has two primary functions, a transcriptional co-activator binding to members of the TCFs/LEF family (TCF1, TCF3, TCF4, or LEF1) to regulate gene expression and a component of the adherens junction regulating cell-to-cell communication [44]. TCFs/LEF are differentially expressed in cells and tissues [52][53][54]. TCF1 (encoded by the TCF7 gene) was the first TCF/ LEF family member discovered along with LEF1 (encoded by the LEF1 gene) [53,55,56]. TCF1 can act as either a transcriptional activator or repressor, depending on cell and tissue type [57][58][59]. LEF1 is generally known as an activator of transcription [56,60]. TCF3 (encoded by the TCF7L1 gene) is the most abundant TCFs/LEF member and is generally known as a repressor for transcription [60][61][62][63][64]. TCF4 (encoded by the TCF7L2 gene) is the most extensively studied member of the TCF/LEF family because it is more ubiquitously expressed in many human adult tissues in comparison to TCF1, TCF3, and LEF1 [65]. Like TCF1, TCF4 can act as either a transcriptional activator or repressor depending on cell and tissue type [58,59,66,67]. Astrocytes are unique in that they express all TCFs/LEF members [44] and robustly express β-catenin [44]. The Wnt/β-catenin signaling pathway is an important neuroprotective pathway and its dysregulation is associated with a number of neurodegenerative diseases, including AD, PD, HAND, and psychiatric disorders such as bipolar disorder and depression [46,68,69]. This pathway is vital to various functions in the CNS ranging from memory consolidation in astrocytes, neurogenesis, neurotransmitter release, induction of long-term potentiation and depolarization resulting in increased synaptic strengths [46,70,71].
We assessed here the direct impact of β-catenin on IL-6 gene expression in astrocytes. We reveal a complex mechanism where by the effect of β-catenin on IL-6 is discordant with TCFs/LEF. Typically, β-catenin complexes with TCFs/LEF to regulate gene expression. Here, we show that β-catenin inhibits IL-6 promoter activity while also inducing NF-κB and C/EBP, two activators of IL-6. TCFs/LEF on the other hand induce IL-6 gene expression. This complex regulation of IL-6 expression in the CNS underscore its tight regulation as high and/or unregulated IL-6 levels can be neuroinflammatory.

Bioinformatic search for TCFs/LEF DNA binding sequences in IL-6 promoter
The locus of IL-6 gene on human genome was identified by searching on PubMed. Approximately 1.4 kb region upstream of transcription start site was identified as putative promoter region. The promoter elements such as CRE, C/EBP, NF-κB, and AP-1 previously characterized for IL-6 [73,74] were all found to be present in the putative promoter sequence. The Sequencher program, software version 5.4.6 (Genes Code incorporation, Ann Arbor, MI) was used to locate TCFs/LEF binding sites on the IL-6 promoter using the TCFs/LEF consensus sequence 5′-CAAAGA-3′.

Statistical analysis
Statistical analyses were performed with consultation of Rush statistical core using Prism software (GraphPad Prism, San Diego, CA). The variables were compared using either the two-tailed one sample T-test or the one-way ANOVA with the post hoc Dunnett's on the data. p-value of ≤0.05 was considered significant and all experiments were performed independently at least three times.

β-Catenin negatively regulates IL-6 transcription in human astrocytes
A previous study from our lab demonstrated that inhibition of β-catenin in human astrocytes leads to astrocyte senescence [41]. Given that IL-6 expression constitutes a Senescence Associated Secretory Phenotype (SASP) [75,76], we assessed whether β-catenin regulates IL-6 gene Fig. 1 β-catenin negatively regulates IL-6 transcription in normal human astrocytes (NHAs). a NHAs were transfected with β-catenin or scrambled (scrm) siRNA and at 48 h, β-catenin and IL-6 mRNA levels were measured by real-time PCR. Efficacy of KD of β-catenin protein is shown by western blot (Ai) and quantified with densitometer (Aii). b NHAs were transfected with β-catenin or scrm siRNA and at 48 h IL-6 protein was measured by ELISA. c Astrocytes were either treated with activators of β-catenin (BIO at 0.5 μM or LiCl at 5 mM) or left untreated and at 24 h, βcatenin protein was measured by Western Blot (Ci) and quantified by densitometry (Cii-Ciii). Effect of β-catenin activators on IL-6 mRNA (d) and IL-6 protein (e) were measured by real-time PCR and ELISA, respectively. *indicates p ≤ 0.05 in comparison to respective control. The data is presented as mean ± SEM (n = ≥3; one sample T-Test) expression. Astrocytes express IL-6 mRNA as determined by qPCR (Supl. Figure 1). siRNA knockdown (KD) of β-catenin induced IL-6 mRNA and protein expression by~3-2-fold, respectively ( Fig. 1ai and b). Efficiency of β-catenin mRNA KD was consistently greater than 90% (Fig. 1aii). Conversely, induction of β-catenin using two pharmacologic agents (BIO and LiCl) [77,78] (Fig. 1c) reduced IL-6 mRNA and protein by~75 and 40%, respectively (Fig. 1d and e). Together, these lossand gain-of-function studies demonstrate that β-catenin negatively regulates IL-6 transcription in astrocytes.

Identification of functional TCFs/LEF binding sites on the IL-6 promoter
Given the discordance between β-catenin and TCFs/LEF effects on IL-6 gene transcription, we probed the molecular mechanism by which TCFs/LEF positively regulates IL-6 mRNA. Using bioinformatics analyses, we identified two putative regions for TCFs/LEF binding (CAAAGA) in the human IL-6 promoter, a proximal site located at -91 nt and a distal site at -948 nt from the transcription start site (Fig. 3a). We then amplified and cloned this entire 1.4 kb putative IL-6 promoter region into the pGL4.12 luciferase reporter vector. To determine whether the insert is a promoter element for luciferase gene and whether this activity was specific to IL-6 gene expression, we transfected human astrocytes with pGL4.12 (p501, vector control) and the IL-6 promoter reporter plasmid (p563B). Twenty-four hour posttransfection, the cells were either stimulated or unstimulated with an inducing agent specific for IL-6 expression (125 ng/mL IL-1β and luciferase activity was measured 24 h post treatment. The IL-6 insert functioned as a promoter for the pGL4.12 vector under basal conditions (Suppl. Fig. 3a) and stimulation with IL-1β significantly induced IL-6 promoter activity (Suppl. Fig.  3b). These studies demonstrate that putative DNA binding sites for TCFs/LEF within IL-6 promoter is functional to drive promoter expression and is inducible (Suppl. Fig. 3a and b). Knockdown of β-catenin in Fig. 3 Identification of functional binding sites of TCFs/LEF and preferential tethering of TCF/LEF family members on IL-6 promoter. a Identification of two putative TCF/LEF sites located between − 91 and − 86 (1st site; proximal) and − 948 and − 943 (2nd site; distal) on the IL-6 promoter by bioinformatic analysis. b and c ChIP was performed using 4 μg of β-catenin, TCF1, TCF3, TCF4, LEF1, or isotype IgG control antibodies and DNA amplified spanning the proximal (b) or distal (c) sites on the IL-6 promoter. *indicates p ≤ 0.05 in comparison to isotype IgG control (n = 5 for all and n = 2 for β-catenin ChIP , one sample T-Test). d Wild-type (wt) IL-6 reporter plasmid was mutated by substitution on the proximal (e), distal (f), or both (g) TCF/LEF putative binding sites. h NHAs were transfected with either the wt IL-6 promoter plasmid or the mutated plasmids shown in (d), (e), and (f). At 48 h post transfection, IL-6 promoter activity was measured through luciferase reporter assay. * indicates p ≤ 0.05 in comparison to IL-6 wt. The data is presented as mean ± SEM (n = 5, one-way ANOVA, post hoc Dunnett's) cultures transfected with the wild type IL-6 reporter plasmid containing sites for TCFs/LEF DNA binding induced IL-6 reporter activity by 1.5-fold (Suppl. Fig. 3c). This data further confirms that β-catenin inhibits IL-6 at the transcriptional level.

Preferential tethering of TCF/LEF family members on IL-6 promoter
We performed chromatin immunoprecipitation (ChIP) assays to assess whether β-catenin and/or TCFs/LEF are tethered on the distal and/or proximal TCFs/LEF sites within the IL-6 promoter. Using antibodies for βcatenin, TCF1, TCF3, TCF4, LEF1 or isotype IgG control, followed by qPCR for endogenous IL-6 promoter region encompassing both proximal and distal TCFs/ LEF binding sites (Fig. 3a), we show that TCF3 and TCF4 are bound at both the proximal and distal TCF/ LEF sites while LEF1 was bound only at the proximal site. β-catenin exhibited strong binding affinity on both sites (10-30 fold, respectively) ( Fig. 3b and c). TCF1 was not bound at either site at a statistically significant level in comparison to IgG.

TCFs/LEF sites at the proximal and distal sites are important for IL-6 promoter activity
To assess whether the proximal and/or distal TCFs/LEF sites (Fig. 3d) on the IL-6 promoter regulates basal and inducible IL-6 activity, we performed site-directed mutagenesis. We generated three constructs with substitution (CAAAGA → CCTGGA) at the proximal site (Fig. 3e), distal site (Fig. 3f), or at both the proximal and distal binding sites (Fig. 3g). Endogenous IL-6 reporter activity was significantly reduced by mutating the distal site, the proximal site, and mutations at both site by~70% (Fig.  3h). These data demonstrate that both proximal and distal TCFs/LEF sites are essential for IL-6 promoter activity.

TCFs/LEF induction of IL-6 is independent of β-catenin and depends on other transcriptional co-regulators (ATF2 and SMADS)
Canonical β-catenin signaling involves the binding of βcatenin to TCFs/LEF family members to regulate gene expression. Our data indicate that β-catenin and TCFs/ LEF have a discordant impact on IL-6 gene regulation. Specially, β-catenin inhibits while TCFs/LEF induces IL-6 transcription. To determine which transcriptional coregulators may be involved in TCFs/LEF induction of IL-6, we assessed the role of Activating Transcription Factor 2 (ATF2), Caenorhabditis elegans SMA ("small" worm phenotype) and Drosophila MAD ("Mothers Against Decapentaplegic") family of genes (SMADs), and Zinc finger E-box binding homeobox 1 (ZEB1) in induction of IL-6. All three can partner with TCFs/LEF to regulate gene expression independent of β-catenin [79][80][81][82][83][84][85]. We show that human astrocytes express ATF2, SMADs, and ZEB1 to levels relative to GAPDH (Suppl. Fig. 4). To determine which of these co-factors may also be regulating IL-6, we either knocked them down individually using siRNAs or treated the cells with small molecule inhibitors. SB525334 is a small molecule inhibitor for Smads 2 and 3 and LDN-193189 is a small molecule inhibitor for Smads 1, 5, and 8/9. Efficacy of KD for ATF2 and ZEB1 was measured by qPCR and was consistently ≥85% (Figs. 4a and b). Efficacy of small molecule inhibitor was assessed by measuring their effect on their respective target genes. A downstream target of Smads 2 and 3 is Plasminogen activation inhibitor (PAI) while inhibitor of DNA binding protein (Id1) is a downstream target of Smads 1, 5, and 8/9 (Figs. 4c and d).

Discussion
Astrocytes secrete a number of cytokines and their regulation is likely to be tightly coordinated and regulated as their over-expression may lead to inflammation in the CNS, which is highly neurotoxic. We assessed here the molecular mechanism by which astrocytes regulate IL-6 gene expression. We show that IL-6 regulation at the transcriptional level is indeed complex. β-catenin partners with TCF/LEF at two newly identified sites on the IL-6 promoter to inhibit IL-6 expression, yet it also induces NF-κB and C/EBP, two factors well established to induce IL-6 promoter activity [35,89,90]. Likewise, TCFs/LEF family members, which typically partner with β-catenin, function independently of β-catenin to induce IL-6 transcription through partnering with ATF2 and SMAD family members. Taken together, these data demonstrate that atypical activity of β-catenin and TCFs/LEF regulate IL-6 gene expression in astrocytes.
Astrocytes express robust level of canonical Wnt/β-catenin signaling and express all of the TCFs/LEF family members, making astrocytes an ideal cell type to study canonical Wnt/β-catenin signaling [43,44,91]. Our working model informed by published [92][93][94] and data presented here indicate that under homeostatic conditions, where there is robust β-catenin expression, βcatenin induces NF-κB and C/EBP to induce IL-6 promoter activity. This induction can be viewed as a "normal" level of IL-6 that is present in the CNS under normal conditions to promote myelination of oligodendrocyte, synaptic transmission, synaptic plasticity, and overall cell survival. However, under nonhomeostatic conditions (e.g. inflammation), β-catenin level is reduced [41-43, 91, 95], as such TCFs/LEF can now partner with factors other than β-catenin. Here, it is ATF2 and members of the SMAD family, to induce IL-6 expression. The level may potentially be higher than what is typically found under physiologic conditions and as such, the increased IL-6 expression can lead to heightened inflammation. High IL-6 level is a hallmark feature of a number of neurodegenerative diseases including MS, AD, PD and brain tumors [20].
Our studies further demonstrate that TCFs/LEF can function independently of β-catenin to regulate gene expression. β-catenin typically partners with one member of TCFs/LEF to regulate gene expression. For example, we previously demonstrated that β-catenin partners with TCF4 to inhibit HIV replication at the transcriptional level in astrocytes [43], it partners with TCF1 to induce Excitatory Amino Acid Transporter 2 (EAAT2) gene expression, and it partners with TCF3 to induce Glutamine Synthetase (GS) gene expression [72]. Here, we show discordant β-catenin and TCFs/LEF effects on IL-6 promoter activity. Interestingly, any of the TCFs/LEF family members can induce IL-6 expression and they may do so as a complex as knocking one member down did not interfere with gene expression of the other members. While all TCF3, TCF4, and LEF1 were found to be tethered on IL-6 promoter either at both proximal and distal site or only at distal site, TCF1 tethering on IL-6 promoter did not reach statistically significant levels beyond IgG isotype control. This could be that its binding is low to be significant, although in some experiments higher binding was trending. Limited studies demonstrate the ability of TCFs/LEF to function as a complex to regulate gene expression. In one study, TCF4 and LEF1 were reported to form a complex to regulate gene expression [96]. More often than not, TCFs/LEF family members exert complementary roles. For example, a TCF1/LEF1 double knockout murine model caused limb deficiencies [57] and a TCF1/TCF4 double knockout murine model led to hindgut defects and a gastrointestinal tract transformation [97].
Much of the molecular mechanism regulating IL-6 at the transcriptional level is based on studies in immune cells. Those studies indicate that C/EBP and NF-κB positively regulate IL-6 [34,36,86,98,99]. One study in astrocytes demonstrated that IL-6 is positively regulated by C/EBP [36]. We also show C/EBP-δ and NF-kB inhibition disturbed IL-6 expression and simultaneously show that β-catenin positively regulates C/EBP-δ and NF-κB. This finding suggests that β-catenin may be a master regulator of IL-6 expression, whereby in presence of its Wnts transcriptional partners, it inhibits promoter activity of IL-6. But in absence of β-catenin, TCFs/LEF, C/EBP, and NF-κB all act independently to activate IL-6 transcription.
Inflammation is not always a negative response. In fact, inflammation is often needed for repair. However, it is hyperinflammation, uncontrolled inflammation that is detrimental. Our data indicate that β-catenin may be a master regulator of NF-κB and C/EBP mediated inflammatory responses. Both NF-κB and C/EBP can be inflammatory through induction of cytokines, which can be either pro or anti-inflammatory in relation to amount and context. As indicated, IL-6 can be either inflammatory or non-inflammatory depending on its signaling pathway and whether it binds to cell associated cognate receptor or soluble receptor. Therefore, we propose that β-catenin is a mean by which to override NF-κB and C/ EBP mediated "over activation" and when β-catenin is diminished, there is an unchecked response to NF-κB Fig. 5 Inhibition of β-catenin downregulates C/EBP and NF-κB activity in NHAs. a NHAs were transfected with C/EBP-β, C/EBP-δ, RelA, or scrm siRNA and at 48 h, their respective mRNA expression was measured by real-time PCR. b NHAs were transfected with C/EBP-β, C/EBP-δ, Rel A, or scrm siRNA and at 24 h co-transfected with C/EBP or NF-κB reporter plasmids and 24 h later luciferase reporter activity was measured. c NHAs were transfected with C/EBP-β, C/EBP-δ, RelA, or scrm siRNA and at 48 h IL-6 mRNA expression was measured by real-time PCR. d NHAs were transfected with β-catenin, TCF1, TCF3, TCF4, LEF1, and scrm siRNA for 24 h then transfected with C/EBP or NF-κB reporter plasmid for another 24 h followed by luciferase reporter activity measurement at 48 h post transfection. * indicates p ≤ 0.05 in comparison to respective control. The data is presented as mean ± SEM (n = ≥3, one-way ANOVA, post hoc Dunnett's) and C/EBP activity in glial cells that can lead to hyperinflammation.
β-catenin and TCF/LEF may inform novel therapeutic strategies to regulate IL-6 levels in the CNS. Tocilizumab, a humanized anti-IL-6 receptor antibody is now approved for the treatment of rheumatoid arthritis and juvenile idiopathic arthritis [100], but it may not exert an effect in reducing or ameliorating high IL-6 in the CNS in context of neurodegenerative diseases due to poor penetration of antibodies into the CNS. IL-6, however, is not always proinflammatory [20,45]. It can also exert anti-inflammatory responses through trans-signaling [20,45]. Therefore, in conditions where IL-6 levels may be needed to be increased within the CNS, to promote repair and survival, inducing TCFs/LEF can be a pathway by which this can occur. In contrast, in conditions where elevated IL-6 is well documented to be associated with neurodegenerative diseases, inhibiting IL-6 may be an approach to reduce and/or ameliorate IL-6 in the CNS. Here, induction of β-catenin may be such an approach. While there are several small molecules that can modulate β-catenin signaling, (https://www. selleckchem.com/Wnt.html), β-catenin signaling is highly context dependent and should be specifically targeted to cell in question. Advances in cell-specific gene therapy may be an approach by which IL-6 expression in the CNS is regulated through β-catenin/ TCFs/LEF manipulation to ensure that there are no off-target effects. Continuing to reveal molecular mechanism(s) by which cytokines in the CNS are regulated is critical to reveal their unique regulation in the CNS that may be different than in other organ systems to inform brain-specific targeted strategy to control cytokine-mediated pro-and anti-inflammatory responses in the CNS.

Conclusions
In summary, we demonstrate a novel and discordant canonical Wnt/β-catenin signaling regulation of IL-6 expression in human astrocytes. Specifically, β-catenin inhibits IL-6 at the transcription level, whereas TCF/LEF partner with other transcription factors to induce IL-6 transcription. β-catenin also regulates NF/κB and c/EBP expression, potentially functioning as a signal that can overcome excessive induction of IL-6 in the brain.