key: cord-0829554-qyq1lkry
authors: Noori, Mahboubeh S.; Courreges, Maria C.; Bergmeier, Stephen C.; McCall, Kelly D.; Goetz, Douglas J.
title: Modulation of LPS-induced inflammatory cytokine production by a novel glycogen synthase kinase-3 inhibitor
date: 2020-07-04
journal: Eur J Pharmacol
DOI: 10.1016/j.ejphar.2020.173340
sha: 927f05006957d1d526d0ff544e0bbd644bba28ca
doc_id: 829554
cord_uid: qyq1lkry

Sepsis is a serious condition that can lead to long-term organ damage and death. At the molecular level, the hallmark of sepsis is the elevated expression of a multitude of potent cytokines, i.e. a cytokine storm. For sepsis involving gram-negative bacteria, macrophages recognize lipopolysaccharide (LPS) shed from the bacteria, activating Toll-like-receptor 4 (TLR4), and triggering a cytokine storm. Glycogen synthase kinase-3 (GSK-3) is a highly active kinase that has been implicated in LPS-induced cytokine production. Thus, compounds that inhibit GSK-3 could be potential therapeutics for sepsis. Our group has recently described a novel and highly selective inhibitor of GSK-3 termed COB-187. In the present study, using THP-1 macrophages, we evaluated the ability of COB-187 to attenuate LPS-induced cytokine production. We found that COB-187 significantly reduced, at the protein and mRNA levels, cytokines induced by LPS (e.g. IL-6, TNF-α, IL-1β, CXCL10, and IFN-β). Further, the data suggest that the inhibition could be due, at least in part, to COB-187 reducing NF-κB (p65/p50) DNA binding activity as well as reducing IRF-3 phosphorylation at Serine 396. Thus, COB-187 appears to be a potent inhibitor of the cytokine storm induced by LPS.

Sepsis is caused by a dysregulated host response to infection (Schulte et al., 2013) and septic shock is a leading cause of death (Schulte et al., 2013; Sharma et al., 2017) . Thus, there is an ongoing effort to develop therapeutics for this devistating condition (Sharma et al., 2017) .

Sepsis is entwined with host-response to invasive pathogens during which the innate immune system recognises pathogen-associated molecular patterns (PAMPs) via pathogen recognation receptors including toll-like receptors (TLRs) (McCoy, 2016; Moresco et al., 2011; Schulte et al., 2013; Takeda and Akira, 2001) . This recognition induces the production and release of inflammatory mediators (e.g. cytokines including a subset termed chemokines) which help the host organism eliminate the invasion (Dinarello, 2000; Martin et al., 2005; O'Neill and Dinarello, 2000) . That said, uncontrolled production of cytokines is thought to be the major driving force in sepsis (Schulte et al., 2013) .

In the case of sepsis due to gram-negative bacteria, uncontrolled activation of one member of the TLR family, namely TLR4, via LPS, a component of gram-negative bacteria, leads to production of inflammatory cytokines through the myeloid differentiation factor 88 (Myd88)dependent and/or -independent pathways (Jope et al., 2016) . Sepsis is often characterized by the presence of excessive levels of pro-inflammatory cytokines (e.g. TNF-α and IL-6), as well as the anti-inflammatory cytokine IL-10, in plasma and/or serum (Chaudhry et al., 2013) . Nuclear factor-kappa-light-chain-enhancer of activated B cells (NF-кB) plays an important role in cytokine regulation and the level of NF-кB activity is strongly corrolated with the severity of sepsis (Arnalich et al., 2000; Chaudhry et al., 2013; Liu and Malik, 2006) . Thus, inhibitng LPSinduced cytokine production is a potential therapeutic approach for gram-negative bacteriainduced sepsis.

GSK-3 appears to play an important role in the LPS-induced production of inflammatory cytokines via the TLR4 pathway. Specifically, in the TLR4 Myd88-dependent pathway, GSK-3 is involved in activation of NF-кB as well as the inhibition of activator protein 1 (AP-1) and cAMP response element binding protein (CREB) transcription factors (Jope et al., 2016; Ko and Lee, 2016) . In the TLR4 Myd88-independent pathway, GSK-3 may help regulate the production of type I IFNs (e.g. IFN-β) through TANK-binding kinase 1 (TBK1) and interferon regulatory factor-3 (IRF-3) (Jope et al., 2016) . Given these observations and the important role of GSK-3 in regulating inflammatory cytokines noted by others (Beurel et al., 2010; Wang et al., 2014 Wang et al., , 2011 , it is not surprising that GSK-3 inhibitors have been touted as potential therapeutics for sepsis (Martin et al., 2005) .

Our group recently described a novel inhibitor of GSK-3 termed COB-187 (Noori et al., 2019) . In molecular assays, COB-187 was found to inhibit both GSK-3α and GSK-3β with an IC 50 in the nM range and to be highly selective for GSK-3 amongst the over 400 kinases screened (Noori et al., 2019) . COB-187 appeared to inhibit cellular GSK-3 as evidenced by its ability to inhibit the phosphorylation of canonical GSK-3 cellular substrates (Noori et al., 2019) .

In the present study we sought to determine if COB-187 could inhibit LPS-induced cytokine storms.

(ATCC; Manassas, VA) were cultured and differentiated to the macrophage phenotype via PMA as described previously (Noori et al., 2019) . The effects of LPS and COB-187 on the metabolic activity of THP-1 macrophages were evaluated using the CellTiter 96 AQ ueous One Solution Cell Proliferation Assay solution (Promega; Madison, WI) according to manufacturer's protocol. thiazolidine-2-thione] was synthesized and purified to >95% purity (Noori et al., 2019) . For compound treatments, stock COB-187 (200 mM in 100% DMSO) was diluted in complete culture medium to the final desired concentrations and 0.1% DMSO (v/v) (Sigma-Aldrich). PMA-differentiated THP-1 macrophages (THP-1 macrophages) were pretreated with varying concentrations of COB-187 or 0.1% DMSO (solvent control) for 1 h prior to 4 h stimulation with 10 ng/ml of LPS (Sigma-Aldrich, E. coli O111:B4 strain) in medium containing the same concentration of COB-187 or 0.1% DMSO alone.

The presence of inflammatory cytokines in culture supernatants was quantified using the proteome profiler human cytokine array kits (a membrane-based sandwich immunoassay) (ARY005B, R&D Systems; Minneapolis, MN). The procedure was as per the manufacturer's instructions. Blots were scanned and analyzed using the Bio-Rad ChemiDoc XRS+ Molecular Imager (Hercules, CA) and the pixel density of the spots on the blots were measured using Quick Spots software (HLImage ++ ; R&D Systems). The measured density of the negative control spot (background) was then subtracted from all other density measurements. Each density result was then normalized to the density of the reference control spots. Human IL-6, IL-10, IFN-β, and CXCL10/IP-10 Quantikine ELISA kits (R&D Systems) were used according to the manufacturer's protocol.

Quantification of Changes in the Gene Expression. In parallel to collecting the supernatant for cytokine analysis at the protein level, the THP-1 macrophages were harvested, and their mRNA was used to evaluate the expression of IL-6, TNF-α, IL-1β, CXCL10, and IFN-β using RT-qPCR as described previously (Noori et al., 2017) . IL-6, TNF-α, IL-1β, CXCL10, IFN-β Taqman® Gene Expression Assays (human, FAM-MGB, Thermo Fisher), Taqman® Gene Expression Master Mix (Applied Biosystems) or SYBR Green Master Mix (Bio-Rad), StepOnePlus™ Real-Time PCR System (Applied Biosystems), and housekeeping genes ACTB and/or GAPDH (human, VIC-MGB, Thermo Fisher) were used to perform the RT-qPCR. ∆∆CT method was utilized for gene expression comparisons relative to the 0.1% DMSO control in combination with LPS.

THP-1 macrophage mRNA transcripts of the TLR signaling pathway were investigated using Human Toll-Like Receptor Signaling Pathway RT² Profiler™ PCR Arrays (SABiosciences; Qiagen). In brief, total RNA was isolated as described previously (Noori et al., 2017) and cDNA was synthesized with the RT 2 First Strand Kit (SABiosciences) using an Eppendorf Mastercycler gradient. RT-qPCR was performed using Human Toll-Like Receptor Signaling Pathway RT² Profiler™ PCR Arrays (SABiosciences), RT² Real-Time™ SYBR Green/ROX PCR Master mix (SABiosciences), and the StepOnePlus™ Real-Time PCR System (Applied Biosystems). Gene expression data was normalized to the reference gene with the most stable expression (ACTB). Web-based SABioscience software was used to analyze the data, compare the results from different treatments, and evaluate P-values.

Note that RNA samples yielding a ratio of absorbance 260 nm/ absorbance 280 nm between 1.8 and 2, and a ratio of absorbance 260 nm/ absorbance 230 nm greater than 1.7 were considered pure. Ribosomal RNA band integrity was also evaluated and samples with RNA Integrity Number (RIN) of 9.5 or higher were used for cDNA synthesis and RT-qPCR.

Assay; Evaluating DNA-Binding Activity. Nuclear and cytoplasmic protein fractions were extracted using Nuclear Extraction kit (Active Motif; Carlsbad, CA) according to the manufacturer's protocol. The protein content was quantified using the micro BCA protein assay kit (Thermo Fisher) and Nanodrop 2000 Micro-volume UV-Vis Spectrophotometer (Thermo Fisher). Five µg of extracted proteins were used in TransAm transcription factor ELISA assays (Active Motif) to measure the DNA binding activity of transcription factors NF-кB and IRF-3. Samples were analyzed using a Synergy HT Multi-Mode Microplate Reader (BioTek).

Phosphorylation of IRF-3 at Serine 396 in 10 µg of extracted nuclear and cytoplasmic proteins from the previous section was evaluated using western blot analysis as described previously (Noori et al., 2019) . Rabbit anti-IRF-3 (D83B9) and anti-phospho-IRF-3 (Serine 396, 4D4G) monoclonal antibodies (Cell Signaling Technology; Danvers, MA) were used to detect the level of total IRF-3 and its phosphorylation at Serine 396 [pIRF3 (S 396 )], respectively. IRDye 680LT and 800CW goat anti-rabbit IgG (H+L) polyclonal secondary antibodies (LI-COR Biosciences) and the LI-COR Odyssey Infrared Imaging System were used for visualization and quantification of signals. Odyssey blocking buffer was used to dilute antibodies. The ratio of pIRF3 (S 396 ) to total IRF-3 was used in the statistical analysis.

Chemokines/Cytokines Involved in Sepsis. A proteome profiler human cytokine array (protein cytokine array) was used to investigate the effect of COB-187 on LPS induction of cytokines in THP-1 macrophages. Representative blots from this analysis are provided in Fig. 1 . As revealed by comparing the top two blots presented in Fig. 1 , nine duplicate pairs of dots are distinctly present in the blot prepared from supernatants of LPS treated THP-1 cells (second blot) that are not, or faintly, present in the blot prepared from supernatants of THP-1 cells not treated with LPS (top blot). These dots were the location of capture antibodies for: CCL3/CCL4, CCL5, CXCL1, CXCL10, IL-1β, IL-1ra, IL-8, MIF, and TNF-α. Although the pixel analysis identified additional cytokines that appeared to be elevated, these additional cytokines were all < 6% of the internal reference control and were difficult to discern visually. In contrast, the nine proteins highlighted in Fig. 1 were all > 45% of the reference standard and were readily discerned visually. Thus, we continued with the analysis of the nine proteins listed in Fig. 1 .

As shown in the four lower blots of Fig. 1 , the pixel density of the nine identified dots appeared to decrease, in a dose dependent manner, when the THP-1 cells were treated with COB-187 concurrent with LPS-stimulation. Replicate analysis was performed and the results from the replicates averaged and presented in Fig. 2 . As shown in Fig. 2A -Fig. 2D , COB-187 treatment with LPS-stimulation resulted in a significant reduction in the LPS-induced production of the chemokines with the least inhibitory effect observed for CCL3/CCL4 ( Fig. 2A) , and the most dramatic effect observed for CXCL10 (Fig. 2D ). Addition of COB-187 caused a significant decrease in the LPS-induced expression of the pro-inflammatory cytokines TNF-α, IL-1β, MIF, and IL-8, as well as the anti-inflammatory cytokine IL-1ra (Fig. 2) . TNF-α and IL-1β were reduced to essentially background levels while the effects on IL-8 and MIF were the less dramatic.

IL-6 and IL-10 are reported to be important in sepsis (Chaudhry et al., 2013; Martin et al., 2005; Schulte et al., 2013) and the protein levels of these cytokines were poorly detected using the cytokine array. Therefore, the Quantikine IL-6 and IL-10 ELISA kits were used to evaluate the expression of IL-6 and IL-10, respectively. As presented in Figs. 3A and 3B, LPS stimulation significantly increased the secretion of both cytokines. Treatment with COB-187 at concentrations ≥ 5 µM significantly reduced the LPS-induced IL-6 production in a dosedependent manner (Fig. 3A) . Interestingly, for IL-10 production ( Fig. 3B) , treatment with the lowest concentration of COB-187 tested (i.e. 2.5 µM) significantly increased the production of anti-inflammatory cytokine IL-10, while COB-187 concentrations higher than 5 µM brought the IL-10 level back to the basal level.

The capture antibody for IFN-β, a cytokine important in sepsis (Mahieu and Libert, 2007) , was not included in the proteome profiler human cytokine array. Therefore, the Quantikine IFNβ ELISA kit was used to evaluate the effect of COB-187 on the expression of IFN-β, one of the type I IFNs regulated by the Myd88-independent pathway (Hoshino et al., 2002) . As shown in Table 1 (column labeled Protein). As shown, the IC 50 s are all in the micromolar range with the lowest value (greatest inhibitory effect) observed for CXCL10 (IC 50 of 2 µM) and the highest value (least inhibitory effect) observed for MIF (IC 50 of 48 µM). The average IC 50 for the cytokines presented in Table 1 was 15 µM with a standard deviation of 16 µM. (Table 2) , with the exception of IFN-α1, IFN-γ and IL-2 (Supp. Table 1 ). The mRNA expression of all the inflammatory cytokines and chemokines that were significantly induced by LPS, were significantly inhibited by treatment with 25 µM COB-187 (Table 2) ; the exception being lymphotoxin alpha which was slightly induced by LPS and not inhibited by COB-187 (Supp. Table 1 1-3, namely CCL3/CCL4, CCL5, CXCL1, IL-1ra, and MIF, were not in the TLR array. Overall, COB-187 has a dramatic inhibitory effect on a host of cytokine and chemokine mRNA transcripts induced by LPS.

CXCL10, IL-6, IL-1β, TNF-α, and IFN-β are among the most important and well-studied chemokines/cytokines involved in sepsis (Bakshi et al., 2017; Chaudhry et al., 2013; Herzig et al., 2014; Jope et al., 2016; Mahieu and Libert, 2007; Schulte et al., 2013) . Therefore, the doseresponse of COB-187 inhibition of these cytokines was further investigated using RT-qPCR ( Fig. 4) . [Note that IL-10 was not further investigated due to its bimodal protein response ( Fig.  3B )]. A best fit regression [i.e. third-degree polynomial] of the data presented in Fig. 4 was performed to determine the IC 50 . The IC 50 for all of the genes was ≤ 10 µM (Table 1, column labeled mRNA). These results are in line with the IC 50 values determined at the protein level (Table 1 , column labeled Protein) and the PCR array data presented in Table 2 . Note that 50 µM of COB-187 effectively returned the LPS-induced expression level of all of the cytokines to basal levels (Fig. 4) .

In addition to the cytokines and chemokines presented in Table 2 , the TLR PCR array revealed that COB-187 significantly inhibited LPS induction of a variety of TLR pathway molecules, including co-stimulatory molecules (e.g. CD80 and CD86), kinases (e.g. IRAK2, MAP2K3, MAP4K4, RIPK2, and TBK1), adapter molecules (e.g. MYD88), and transcription factors (e.g. IRF1, REL, RELA, NFKB1, NFKB2, NFKBIA) (Table 3) . Overall, 55 of the 84 genes in the TLR PCR array were induced by LPS and of these 38 (~69%) were significantly reduced by COB-187; 14 (~25%) to background levels (Table 3 and Supp. Table 1 ). The level of gene induction by LPS is broad ranging from a 1.27 to a 21673-fold increase. Interestingly, the most remarkable effect of LPS-induced gene expression occurs with cytokines/chemokines where inductions between two and three orders of magnitude are common (Table 2 ). In contrast, for the other genes, with the exception of CD80 and IRAK2, the induction never reaches two orders of magnitude (Table 3 ).

The expression of five genes (HSPA1A, HSPD1, JUN, MAPK8IP3, and TLR9) was significantly increased upon concurrent treatment with 25 µM COB-187 and LPS relative to LPS treatment alone but these increases were quite modest (ranging from 1.4 to 2.7 fold increase) with the exception of HSPA1A which was increased 32.9 fold with addition of COB-187 (Table   3 ). We did observe 9 genes (ECSIT, FOS, HRAS, IRAK1, LY96, NFKBIL1, SIGIRR, TLR10 and TOLLIP) that were not induced by LPS but that did exhibit an increase in expression upon concurrent treatment with COB-187 and LPS relative to DMSO with LPS or DMSO alone suggesting that COB-187 may induce these genes (Supp. Table 2 ). Again, these effects were relatively small; averaging less than an order of magnitude (Supp. Table 2 ).

Macrophages. GSK-3 has been implicated in the production of pro-inflammatory cytokines, chemokines, and IFN-β in the TLR4 pathway through direct activation of NF-кB (Jope et al., 2016; Ko and Lee, 2016) and /or indirect activation of IRF-3 (Jope et al., 2016) , respectively.

GSK-3 has also been implicated in the production of anti-inflammatory cytokines through inhibition of transcription factors AP-1 and CREB (Jope et al., 2016; Ko and Lee, 2016) . As shown in Figs. 1 -4 and Tables 1 and 2, COB-187 significantly reduced the expression of proinflammatory cytokines and chemokines in the LPS-stimulated THP-1 macrophages. Since the amount of secreted IL-10, an anti-inflammatory cytokine, was relatively low (< 20 pg/ml, Fig.   3B ) compared to the other cytokines (e.g. IL-6 was ~600 pg/ml, Fig. 3A ), we focused on evaluating the effect of COB-187 on the activity of NF-кB and IRF-3, which are involved in the production of pro-inflammatory cytokines and IFN-β, rather than CREB and AP-1, which are mainly involved in the production of anti-inflammatory cytokines like IL-10. Note that NF-кB and IRF-3 are in the Myd88-dependent and Myd88-independent TLR4 signaling pathways, respectively.

The NF-кB transcription factor family consists of five subunits namely p50 (NF-кB1), p52 (NF-кB2), p65 (RelA), RelB, and c-Rel (Tergaonkar, 2006) . Note that high activity of the p65/p50 NF-кB dimer is strongly corrolated with the severity and mortality of sepsis (Arnalich et al., 2000; Chaudhry et al., 2013; Liu and Malik, 2006) . As shown in Fig. 5 , 4-h stimulation with 10 ng/ml of LPS increased the DNA binding activity of p50 and p65 subunits of NF-κB more than 70% in THP-1 macrophages relative to control. Concurrent treatment with COB-187 at concentrations ≥ 25 µM significantly reduced the DNA binding activity of p50 (Fig. 5A) Consistent with these expectations, pIRF-3 (S 396 ) was detected, but only in the nuclear fraction of 4-h LPS-stimulated THP-1 macrophages (Fig. 6 ). Concurrent treatment with 50 µM of COB-187 significantly reduced the level of pIRF-3 (S 396 ) found in the nucleus of LPS-stimulated THP-1 macrophages (Fig. 6) . Somewhat surprisingly given this observation, no significant effect of COB-187 on IRF-3 DNA binding activity was detected in LPS-stimulated THP-1 macrophages using the IRF-3 TransAm assay (Supp. Fig. 1 ). This latter observation could be due to insensitivity of the TransAm assay [note that treatment with LPS resulted in only a ~50% increase in the DNA binding activity of IRF-3 relative to control (Supp. Fig. 1) ]. Setting that caveat aside, the results presented in Fig. 6 suggest that COB-187 inhibits LPS-induced IRF-3 activation. 

Several studies have demonstrated the efficacy of GSK-3 inhibitors in models of endotoxic shock (Dugo et al., 2017; Martin et al., 2005; Noh et al., 2012) . Further, in aggregate, numerous reports indicate that GSK-3 inhibitors abate sepsis through attenuation of LPS-induced cytokine/chemokine production (Gao et al., 2018; Huang et al., 2009; Jing et al., 2004; Kontzias et al., 2012; Martin et al., 2005; Steinbrecher et al., 2005) . The present study revealed that COB-187, a novel GSK-3 inhibitor (Noori et al., 2019) , significantly attenuatets the expression of a plethora of LPS-inducible genes. Specifically, LPS induced 55 of the 84 genes in a TLR array of which 69% of those induced were significantly inhibited by COB-187, and 25% were reduced to basal levels. Cytokines/chemokines known to be induced by LPS and important in sepsis (Bakshi et al., 2017; Chaudhry et al., 2013; Huang et al., 2009; Jope et al., 2016; Martin et al., 2005; Schulte et al., 2013; Tay et al., 2012) , were inhibited by COB-187 with relatively high potency (Table 1) . While in the majority of our studies COB-187 was added pre-LPS challenge, treatment with COB-187 post-LPS stimulation did cause a significant reduction in induced CXCL10 mRNA (Fig. 7) .

These results are in line with the previous studies utilizing other GSK-3 inhibitors [e.g. (Gao et al., 2018; Huang et al., 2009; Jing et al., 2004; Kontzias et al., 2012; Martin et al., 2005; Rehani et al., 2009; Steinbrecher et al., 2005) ] and demonstrate the ability of COB-187 to attenuate the cytokine storm induced by LPS. The inhibitory effect of COB-187 does not appear to be due to a non-specific toxic effect since COB-187 had no statistical effect on THP-1 macrophage metabolic activity (Supp. Fig. 2 ). In regards to the anti-inflammatory cytokine IL-10, previous studies found that GSK-3 inhibitors increased the production of IL-10 (Huang et al., 2009; Martin et al., 2005; Noh et al., 2012; Tay et al., 2012) . A similar result was observed in the present study for LPS-stimulated THP-1 macrophages treated with 2.5 µM of COB-187 (Fig.   3B ). However, at ≥ 10 µM of COB-187, the IL-10 protein level was reduced to basal level and treatment with 25 µM COB-187 significantly reduced LPS-induced IL-10 mRNA (Table 2) .

GSK-3 may regulate IL-10 production, in part, through CREB via the Myd88-dependent pathway (Ko and Lee, 2016) . We have no data on the effect of COB-187 on CREB.

Interestingly, the THP-1 macrophages secreted relatively low amounts of IL-10 which could be due to the short incubation time (4 h) of THP-1 macrophages with LPS.

We gained insights into the effect of COB-187 on the activity of the "pro-inflammatory" NF-кB and IRF-3 transcription factors in the TLR4 Myd88-dependent and Myd88-independent pathways, respectively. LPS ligation of TLR4 induces the activation of NF-кB in the Myd88dependent pathway (Ansaldi et al., 2011; Takada et al., 2004) and the elevated activity of NF-кB after bacterial infection is strongly correlated with the severity of sepsis (Arnalich et al., 2000; Chaudhry et al., 2013; Liu and Malik, 2006) . In line with this observation, we found that LPS dramatically induced the DNA binding activity of p50 and p65 subunits of NF-κB (Fig. 5) .

Takada et al. (Takada et al., 2004) reported a complete suppression of LPS-induced NF-кB activity in GSK-3β gene-deleted cells, and demonstrated the pivotal role of GSK-3 in activation of NF-кB (Takada et al., 2004) . These obervations suggest that inhibition of GSK-3 could reduce the activation of NF-кB and consequently abate symptoms associated with sepsis (Jope et al., 2016; Klamer et al., 2010; Ko and Lee, 2016) ; a conjecture that has been bolstered by numerous reports (Ansaldi et al., 2011; Kotliarova et al., 2008; Wilson and Baldwin, 2008) . In the present study, we observed that treatment with COB-187 at ≥ 25 µM significantly reduced the LPSinduced DNA binding activity of p50 and p65 (Fig. 5) . Furthermore, treatment with 25 µM of COB-187 significantly down-regulated the LPS-induced mRNA level of p50 (NF-кB1), p65

(RELA) and Myd88 in THP-1 macrophages (Table 3 ). These findings suggest that COB-187 inhibition of LPS-induced cytokine expression via the Myd88-dependent pathway may be due, in part, to the inhibition of the NF-κB DNA binding activity (Fig. 5) as well as reduction in mRNA expression of the adapter molecules Myd88, and p50 and p65 subunits of NF-κB (Table 3) .

LPS stimulation also activates the Myd88-independent pathway which trigers IFN-β production through the sequential events of TRIF recruitment, TBK1 activation, IRF-3 phosphorylation at Serine 396, IRF-3 dimerization and translocation to the nucleus, and subsequent IRF-3 binding to the IFN-β gene promoter (Bakshi et al., 2017; Jope et al., 2016) . As shown in Figs. 3C and 4E, COB-187 at ≥ 10 µM reduces IFN-β expression, while the phosphorylation of IRF-3 at Serine 396 was only inhibited at 50 µM of COB-187 (Fig. 6) . The effect of COB-187 on LPS-induced IRF-3 activation is further clouded by our finding that COB-187 did not alter LPS-induced IRF-3 DNA binding (Supp. Fig. 1 ). These somewhat paradoxical observations could be explained by the fact that activated NF-κB and c-Jun are required to bind with IRF-3 on the IFN-β gene enhancer/promoter region to induce the expression of IFN-β1 gene (Fitzgerald et al., 2003; Jin et al., 2014; Kawai and Akira, 2007; Núñez Miguel et al., 2007; Sin et al., 2012) . While we did not investigate c-Jun DNA binding activity, we did observe that LPSinduced NF-κB activity was inhibited at COB-187 ≥ 25 µM (Fig. 5) .

In addition to IFN-β, expression of CCL5 and CXCL10 have been reported to be regulated through the Myd88-independent pathway (Bandow et al., 2012; Yoo et al., 2014) . Further, LPS induction of co-stimulatory surface molecules (e.g. CD80 and CD86) is reported to be mediated through IFN-β (Hoebe et al., 2003; Mahieu and Libert, 2007) . In the present study we observed that COB-187 significantly reduced (i) the protein levels of CCL5 (Fig. 2B) , CXCL10 (Fig. 2D) , and IFN-β (Fig. 3C) , (ii) the mRNA expression levels of CXCL10 (Table 2 and Fig. 4D ), TBK1 (Table 3) , and IFN-β1 (Fig. 4E) , CD80 and CD86 (Table 3) , and (iii) the phosphorylation level of IRF-3 at Serine 396 (Fig. 6) . Thus, COB-187 may inhibit, at least in part, the TLR4 Myd88independent pathway.

As mentioned above, numerous studies have reported efficacy of GSK-3 inhibitors in animal models of sepsis raising the potential for a clinical use of GSK-3 for sepsis. Several GSK-3 inhibitors are being investigated in clinical trials, but the majority of these focus on neurological disorders or cancer. In considering GSK-3 inhibitors for sepsis, it is important to appreciate that the translation of promising reagents to the clinic for sepsis is froth with disappointment. Indeed, the 2019 Final Report from the NIH Working Group on Sepsis ("NAGMSC Working Group on Sepsis," 2019) emphasized that there are no accepted treatments for sepsis despite decades of work and dozens of clinical trials. We have found that COB-187 does sharply attenuate LPS-induced cytokine production and is quite selective in molecular kinase screens relative to the most clinically advanced GSK-3 inhibitor Tideglusib (Noori et al., 2019) . The selectivity of COB-187 may ultimately make COB-187 a better clinical candidate. In this regard, studies that compare the effect of established GSK-3 inhibitors and COB-187 on LPS-induced cytokine expression would be insightful.

Finally, we would like to remark on four unresolved issues raised by the study. First, we infer that the inhibition of cytokine induction is due to COB-187 inhibiting GSK-3. This inference is based on (i) our finding that COB-187 is a selective and potent inhibitor of GSK-3 (Noori et al., 2019) and (ii) previous work by others demonstrating GSK-3 inhibitors diminish LPS-induced cytokine expression (Gao et al., 2018; Huang et al., 2009; Jing et al., 2004; Kontzias et al., 2012; Martin et al., 2005; Steinbrecher et al., 2005) . Again, comparing multiple GSK-3 inhibitors with COB-187 would be insightful. Second, the studies were done with THP-1 macrophages at a short time point. Complementary studies with freshly isolated cells, that consider short-term and late-phase proteins are clearly needed. Third, given that GSK-3 has a multitude of substrates (Cormier and Woodgett, 2017; Frame and Cohen, 2001) , and plays a key role in numerous physiological processes (e.g. glycogen metabolism) (Beurel et al., 2015; Frame and Cohen, 2001; Jope et al., 2007; Saraswati et al., 2018) , GSK-3 inhibition could cause numerous side effects. That said, GSK-3 inhibitors have progressed through clinical trials (del Ser et al., 2013; Hoglinger et al., 2012) and lithium, a known GSK-3 inhibitor, is used clinically.

Finally, several of the mechanisms of viral-related cytokine storms are similar to bacterialrelated mechanisms [e.g. overlap in TLR4 and TLR3 (viral response) signaling pathways (Uematsu and Akira, 2007) ]. Given the current COVID-19 epidemic and the fact that cytokine storms are often a complication of coronavirus infections (Channappanavar and Perlman, 2017; Fung et al., 2020) , it would be quite timely to probe the effect of COB-187 on viral-related cytokine storms.

In summary, COB-187 inhibits a plethora of LPS-induced, THP-1 macrophage expressed, TLR-related genes, demonstrating that COB-187 is a potent inhibitor of LPS-induced cytokine storms. Figures   Fig. 1 . Blots from the proteome profiler human cytokine protein array reveal that nine chemokines/cytokines are markedly induced by LPS and their expression is reduced by COB-187. THP-1 macrophages were stimulated with 10 ng/ml of LPS in combination with varying concentrations of COB-187 or 0.1% DMSO (solvent control). LPS treatment was for 4 h and COB-187 was added 1 h prior to treatment with LPS and was maintained throughout the LPS treatment. Separate proteome profiler blots were exposed to supernatants harvested from the various treatment groups (the treatment groups are labeled to the left of the blot). The four dots in the left most region of each blot and the two dots in the right most region of each blot are the internal reference controls and are used to normalize the data obtained from each blot. Note that each dot is paired, i.e. they are replicates. The numbers demarcate the proteins that appear in response to LPS treatment. These proteins are listed in the table to the right. Inset shows the structure of COB-187 (Noori et al., 2019) . Table 2 . COB-187 significantly reduces LPS-induced mRNA expression of cytokines in the TLR pathway. The effect of LPS and concurrent treatment with COB-187 on the mRNA expression levels of a host of cytokines and chemokines in THP-1 macrophages were evaluated using Human Toll-Like Receptor Signaling Pathway RT² Profiler™ PCR Array in three independent experiments. Web-based SABioscience software was used to analyze the data and determine P-values. LPS treatment was for 4 h and COB-187 or 0.1% DMSO (carrier control) was added 1 h prior to treatment with LPS and was maintained throughout the LPS treatment. Results of LPS stimulation in combination with 0.1% DMSO were compared with results of the 0.1% DMSO control, and the corresponding fold changes and P-values for the cytokine genes in the TLR pathway are presented in the second and third columns. Molecules with fold change ≥ 1.0 and P-value < 0.05 are considered to be significantly induced by LPS and are presented in this table. Results of the LPS stimulation concurrent with 25 µM COB-187 were compared to: (i) the results of LPS stimulation in combination with 0.1% DMSO (fourth and fifth columns), and (ii) the 0.1% DMSO control (sixth and seventh columns). For the fourth and fifth columns, molecules with fold change ≤ 1.0 and P-value ≤ 0.05 are considered to be significantly inhibited by COB-187. *Addition of 25 µM of COB-187 to the LPS treatment resulted in a cytokine level statistically identical to treatment with 0.1% DMSO alone. Figure 6 . 

Imaging Pulmonary NF-kappaB Activation and Therapeutic Effects of MLN120B and TDZD-8

Predictive Value of Nuclear Factor κB Activity and Plasma Cytokine Levels in Patients with Sepsis

Identification of TBK1 complexes required for the phosphorylation of IRF3 and the production of interferon β

LPS-induced chemokine expression in both MyD88-dependent and -independent manners is regulated by Cot/Tpl2-ERK axis in macrophages

Glycogen synthase kinase-3 (GSK3): Regulation, actions, and diseases

Innate and adaptive immune responses regulated by glycogen synthase kinase-3 (GSK3)

Pathogenic human coronavirus infections: causes and consequences of cytokine storm and immunopathology

Role of Cytokines as a Double-edged Sword in Sepsis

Recent advances in understanding the cellular roles of GSK-3. F1000Research 6

Treatment of Alzheimer's Disease with the GSK-3 Inhibitor Tideglusib: A Pilot Study

Gsk-3beta inhibitors attenuate the organ injury/dysfunction caused by endotoxemia in the rat

LPS-TLR4 Signaling to IRF-3/7 and NF-κB Involves the Toll Adapters TRAM and TRIF

GSK3 takes centre stage more than 20 years after its discovery

A tug-of-war between severe acute respiratory syndrome coronavirus 2 and host antiviral defence: lessons from other pathogenic viruses

Discovery and antiinflammatory evaluation of benzothiazepinones (BTZs) as novel non-ATP competitive inhibitors of glycogen synthase kinase-3β (GSK-3β)

The role of CXCL10 in the pathogenesis of experimental septic shock

Upregulation of costimulatory molecules induced by lipopolysaccharide and double-stranded RNA occurs by Trif-dependent and Trif-independent pathways

Tideglusib, a GSK-3 inhibitor, reduces progression of brain atrophy in progressive supranuclear palsy

Differential involvement of IFN-β in Toll-like receptor-stimulated dendritic cell activation

Glycogen synthase kinase-3 negatively regulates anti-inflammatory interleukin-10 for lipopolysaccharide-induced iNOS/NO biosynthesis and RANTES production in microglial cells

Noncanonical NF-κB Pathway Controls the Production of Type I Interferons in Antiviral Innate Immunity

A Novel Signaling Pathway Mediates the Inhibition of CCL3/4 Expression by Prostaglandin E2

Stressed and Inflamed, Can GSK3 Be Blamed?

Glycogen Synthase Kinase-3 (GSK3): Inflammation, Diseases, and Therapeutics

Signaling to NF-κB by Toll-like receptors

Using Small Molecule GSK3β Inhibitors to Treat Inflammation

Glycogen synthase kinase 3β in Toll-like receptor signaling

Kinase inhibitors in the treatment of immune-mediated disease

Glycogen Synthase Kinase 3 inhibition Induces Glioma Cell Death through c-MYC, NF-κB and Glucose Regulation

NF-κB activation as a pathological mechanism of septic shock and inflammation

Should We Inhibit Type I Interferons in Sepsis?

Toll-like receptor-mediated cytokine production is differentially regulated by glycogen synthase kinase 3

Toll-like receptors: practice and methods

Toll-like receptors

Protein Kinase C δ (PKCδ)-Extracellular Signal-regulated Kinase 1/2 (ERK1/2) Signaling Cascade Regulates Glycogen Synthase Kinase-3 (GSK-3) Inhibition-mediated Interleukin-10 (IL-10) Expression in Lipopolysaccharide (LPS)-induced Endotoxemia

Identification of a Novel Selective and Potent Inhibitor of Glycogen Synthase Kinase-3

Phenylmethimazole and a thiazole derivative of phenylmethimazole inhibit IL-6 expression by triple negative breast cancer cells

A Dimer of the Toll-Like Receptor 4 Cytoplasmic Domain Provides a Specific Scaffold for the Recruitment of Signalling Adaptor Proteins

The IL-1 receptor/toll-like receptor superfamily: crucial receptors for inflammation and host defense

Toll-Like Receptor-Mediated Production of IL-1Ra Is Negatively Regulated by GSK3 via the MAPK ERK1/2

Glycogen synthase kinase-3 and its inhibitors: Potential target for various therapeutic conditions

Cytokines in Sepsis: Potent Immunoregulators and Potential Therapeutic Targets-An Updated View

Mitigation of sepsis-induced inflammatory responses and organ injury through targeting Wnt/β-catenin signaling

Activation and regulation of interferon-β in immune responses

Glycogen Synthase Kinase 3β Functions To Specify Gene-Specific, NF-κB-Dependent Transcription

Genetic Deletion of Glycogen Synthase Kinase-3β Abrogates Activation of IκBα Kinase, JNK, Akt, and p44/p42 MAPK but Potentiates Apoptosis Induced by Tumor Necrosis Factor

Roles of Toll-like receptors in innate immune responses

Glycogen synthase kinase-3β inhibition improved survivability of mice infected with Burkholderia pseudomallei

NFκB pathway: A good signaling paradigm and therapeutic target

Toll-like receptors and Type I interferons

Glycogen Synthase Kinase 3: A Point of Convergence for the Host Inflammatory Response

GSK3β and the control of infectious bacterial diseases

Maintenance of Constitutive IκB Kinase Activity by Glycogen Synthase Kinase-3 α/β in Pancreatic Cancer

Interferon β protects against lethal endotoxic and septic shock through SIRT1 upregulation

Target

Profiler™ PCR Array in three independent experiments. Web-based SABioscience software was used to analyze the data and determine P-values. LPS treatment was for 4 h and COB-187 or 0.1% DMSO (carrier control) was added 1 h prior to treatment with LPS and was maintained throughout the LPS treatment. Results of LPS stimulation in combination with 0.1% DMSO were compared with results of the 0.1% DMSO control and the corresponding fold changes and P-values for the TLR-related genes are presented in the second and third columns. Molecules with fold change ≥ 1.0 and P-value < 0.05 are considered to be significantly induced by LPS and are presented in this table. Results of the LPS stimulation concurrent with 25 µM COB-187 were compared to: (i) the results of LPS stimulation in combination with 0.1% DMSO (fourth and fifth columns), and (ii) the 0.1% DMSO control (sixth and seventh columns). Bold text demarcates genes that were induced by LPS and significantly reduced by COB-187, i.e. fold change ≤ 1.0 and P-value ≤ 0.05 in the fourth and fifth columns. *Addition of 25 µM of COB-187 to the LPS treatment resulted in an expression level statistically identical to treatment with DMSO alone.