INH14, a Small‐Molecule Urea Derivative, Inhibits the IKKα/β‐Dependent TLR Inflammatory Response

Abstract N‐(4‐Ethylphenyl)‐N′‐phenylurea (INH14) is a fragment‐like compound that inhibits the toll‐like receptor 2 (TLR2)‐mediated inflammatory activity and other inflammatory pathways (i.e., TLR4, TNF‐R and IL‐1R). In this study, we determined the molecular target of INH14. Overexpression of proteins that are part of the TLR2 pathway in cells treated with INH14 indicated that the target lay downstream of the complex TAK1/TAB1. Immunoblot assays showed that INH14 decreased IkBα degradation in cells activated by lipopeptide (TLR2 ligand). These data indicated the kinases IKKα and/or IKKβ as the targets of INH14, which was confirmed with kinase assays (IC50 IKKα=8.97 μm; IC50 IKKβ=3.59 μm). Furthermore, in vivo experiments showed that INH14 decreased TNFα formed after lipopeptide‐induced inflammation, and treatment of ovarian cancer cells with INH14 led to a reduction of NF‐kB constitutive activity and a reduction in the wound‐closing ability of these cells. These results demonstrate that INH14 decreases NF‐kB activation through the inhibition of IKKs. Optimization of INH14 could lead to potent inhibitors of IKKs that might be used as antiinflammatory drugs.


Introduction
To ll-like receptors (TLRs)a re the most thoroughly studied innate immune receptors. They recognize pathogen-associated molecular patterns (PAMPs), which are molecular structures conserved among different pathogens. [1] They also recognize host-derived molecules nameda larmins (danger-associated molecular patterns (DAMPSs)), which mediates terile inflammation after trauma, stress, and injury. [2] Ligation of PAMPs and DAMPsb yT LRs in antigen-presenting cells, such as macrophages or dendriticc ells, leads to the activation of transcription factors (e.g.,n uclear factor-kappa B( NF-kB) or activator protein 1( AP-1)). [3] These transcription factors promote the production of cytokines and chemokines. These, in turn, can lead to the activation of the adaptive immune system. [4] It has been suggested that the endogenousl igand-recognition response couldb et he underlying mechanism of inflammatory processes observed in autoimmune diseases (e.g.,s ystemic lupus erythematosus, psoriasis) and cancer. [5] Therefore, it might be of advantage to use small-molecule inhibitors to block TLR activity prior to the manifestation of chronic inflammation. [6] In this regard,p revious studies have demonstrated that the selection of small-molecules [7] and peptides [8] that inhibit the inflammatory responsem ediated by several TLRs might be successfully accomplished.
The transcription factor NF-kB is involved in the expression of proinflammatory genes (e.g.,c ytokines and chemokines), and therefore, is am aster regulator of inflammation. NF-kB is activated in response to variouss timuli,s uch as infections and stress signals. [9] Its dysregulation is associated with av ariety of diseases,s uch as atherosclerosis, infections, and cancer progression. [10] Activating signals for TLR or TNF-R receptors lead to phosphorylation in specifics erine residues in the activation loop of the IkB (inhibitor of NF-kB)-kinasesI KKa and IKKb. [11] In turn, the active IKK complexp hosphorylates IkBa at Ser32 and Ser36, which leads to its degradation by ubiquitin-activating enzymes and to the liberation of NF-kB. [12] The cytoplasmic NF-kB then translocates into the nucleus to initiate the expression of over 500 genes involved in inflammation, carcinogenesis, and apoptosis. [13] Different triggers lead to either the activation of the NF-kB canonical pathway (NF-kB1 (p50/105)), which is mediated by IKKa/b/g,o rt he noncanonical pathway mediated by IKKa,w hichl eads to activation of NF-kB2 (p52/p100). [14] A substantial number of small-molecule inhibitors of IKKa and IKKb have been reported,t od ate. These can be classified as adenosine triphosphate (ATP) analogues, allosteric effectors, N-(4-Ethylphenyl)-N'-phenylurea (INH14) is af ragment-like compound that inhibits the toll-like receptor 2( TLR2)-mediated inflammatory activity ando ther inflammatoryp athways (i.e., TLR4, TNF-R and IL-1R and compounds that interact with the Cys179 residue in the activation loop of IKKb. [13] However,o nly af ew of these inhibitors are selective for either of these two kinases. The design of selectives mall-molecule inhibitors of IKKa and IKKb has provent ob ec hallenging because the active sites of both enzymes share high structural homology. [15] In previous studies, we reported ac ollection of small molecules (1-8;s ubsequently namedA T1-AT8 for antagonist) that antagonized TLR2 activity in human cells:H EK293c ellso verexpressing TLR2 (HEK293-TLR2)a nd primary monocytes. [7a] As part of these screening efforts, INH14 (inhibitor14: N-(4-ethylphenyl)-N'-phenylurea;F igure 1A)w as identified as an inhibitor of TLR2-mediated NF-kB activation. However,p reliminary data indicatedt hat the inhibitorya ctivity of the compound on TLR2 was not linked to direct interaction with that protein,a s was the case for AT1-AT8, forwhich reasoni tw as not reported in our previous work.
The purpose of the current study was to unravel the mechanism of inhibition of INH14 downstream of TLR2. To achieve this aim, we utilized transcriptional assays to identify potential target proteins and employed target-based assays for confirmation. We alsod erived the likely bindingm ode of INH14 for the kinase IKKb.I na ddition, we conducted in vivo experiments to evaluatet he anti-inflammatory effect of INH14. Importantly, INH14 decreased the NF-kB constitutive activity in ovarian cancer cells. Overall, this makes INH14 ap romising starting point for the development of potent and selectivei nhibitors of these central kinases.

Results
Inhibition of TLR2-dependent NF-kB activation by INH14 INH14 ( Figure 1A)i safragment-like compound with am olecular weight of only 240 Da. It mainly consists of ab iaryl urea scaffold that is common to an umber of kinase inhibitors;i n particular, compounds addressing the epidermal growth factor receptor( EGFR) [16] or the vascular endothelial growth factor receptor 2(VEGFR-2). [16a, b, 17] To confirmt hat the inhibitory activity of INH14 on TLR2 signaling was dose dependent,w et ransfected HEK293-TLR2c ells with al uciferase reporter tandem: the NF-kB-dependent reporterp lasmid Elam.luc and the constitutivelya ctive Renilla plasmid( to normalize fort ransfection efficiency).A fter incubation of the cellsw ith increasing concentrations of INH14, these were stimulated with the TLR2 ligand triacylated lipopeptide Pam3CSK4 (P3). Chemiluminescence measurements indicated that INH14 reduced TLR2-mediated NF-kB activity in ad osedependentm anner,w ith ah alf-inhibitory concentration of 4.127 mm ( Figure 1B). We obtained parallelr esultsi ft he cells were stimulated with diacylated lipopeptide Pam2CSK4 (P2; not shown).

Inhibition of TNFa and IL-1 signaling by INH14
To assess the selectivity of INH14 forT LR2 inhibition, we tested the effect of INH14 in the activity of two receptors related to TLR signaling.I L-1R shares as imilars ignaling pathway (MyD88dependent) and TNF-R,a lthough working through different upstream signalingc omponents,c onverges at the level of the complex formed by TAK1/TAB1 and downstream proteins. [18] HEK293 cells were transfected with Elam.luc and Renilla plasmids, as described above. The cells were treated for 1h with INH14 and then stimulated with TNFa or IL-1b.U nexpectedly, INH14 inhibited NF-kB activation obtained in both cases (Figure 1C,D ). Thus, we hypothesized that INH14 might be ac ellpermeable small molecule that could interfere with the signaling downstream of TLR2, IL-1R,and TNF-R.

Inhibition of TNFa production by INH14 followingTLR2 or TLR4 stimulation
To furtheri nvestigate whether INH14d ecreased the TLR2mediated proinflammatory activity,w et ested the capacity of the compound to inhibith uman and mouse TNFa production after TLR2 stimulation. Human primary monocytes or mouse RAW264.7m acrophages were incubated with INH14, and then stimulated with P3. The amount of TNFa secreted in the supernatant was quantified by means of ELISA. As shown in Figure 2A,I NH14 reduced TNFa production by mouse macrophages after P3 stimulation from (837 AE 30.28) to (496.6 AE 50.69) pg mL À1 .T he inhibitory effect in TLR2 was more pronounced than that obtained after stimulation of TLR4 with LPS (from (1411 AE 214.3) to (892.8 AE 84.71)pgmL À1 ). The reduction in TNFa productionb yo ther MyD88-dependent TLRs after treatment with INH14 has also been confirmed (i.e.,T LR7/8, TLR5;F igure S1 in the SupportingI nformation).A dditionally, we found ad ecrease in the production of TNFa by human monocytes after INH14 treatment if they were stimulated with P3 (reduced to 60.

Locatingthe target of INH14 downstream of TLR2
Our primary hypothesisw as that INH14 was an antagonist of TLR2. However, our results indicatedt hat INH14 inhibited not only TLR2 activity,b ut also TLR2-related (TLR4 and IL-1R) and unrelated pathways (TNF-R). Although we have not determined the permeability coefficient of INH14, our results indicate that it might be ac ell-permeable molecule. Thus, we explored at whichl evel of the mentioned pathways the compound was effective.O verexpression of proteins described to be downstream of TLRs leads to NF-kB activation independently of TLR stimulation. [19] Hence,w et ransiently transfected HEK293c ells with plasmids encoding the TLRa daptor TIRAP/ Mal, Elam.luc,a nd Renilla. Then, we incubated the cells with INH14 for 5h and measured the luciferase activity.A ss hown in Figure 3A,I NH14 decreased the NF-kB activity attained with Mal expression in ad ose-dependentm anner.L ikewise, we transfected HEK293 cells with ap lasmide ncoding MyD88, the next adaptor downstream of TLR2. Figure 3A shows that INH14 also decreased the NF-kB activity induced by MyD88.
The E-selectin promotor in the Elam.luc reporter plasmid consists of three NF-kB and two AP-1 binding sites. [21] There-  fore, we next analyzed the effect of INH14 in independenta ctivation of both transcription factors. HEK293-TLR2 cells were transfected with the reporter plasmid kb3.luc (whichc ontains three NF-kB binding sites in the luciferase promoter), and then they were stimulated with P2 or P3 ( Figure 4A). In both cases, the NF-kB activity obtained after TLR2 stimulation was reduced if the cells were treated with INH14. Next, we transfected HEK293-TLR2 with an AP1-dependent luciferase reporter plasmid (AP1.luc). Incubation of the cells with INH14b efore treatment with P2 did not reducet he AP-1 transcriptional activity ( Figure 4A). Due to the observed inhibition of IKKa/b activation of NF-kB by INH14 ( Figure 3B), and the inhibition of NF-kB activity (but not AP-1) following TLR2 activation, we postulated that INH14e xerted its inhibitory effect at the level of IKKa/b.
To confirmo ur hypothesis, we monitored the activation of proteins downstream of MAPKKs or IKKa/b.T od os o, HEK293-TLR2 cells were treated with INH14 and then stimulated with P3. Immunoblotting analysis showedt hat the phosphorylation of p38 and JNK was not affected by INH14 treatment, whereas IkBa degradationw as significantly reduced( Figure4B). This result further indicated IKKa/b as the possible targeto fI NH14.
We next proceeded to discard an effect of INH14 in other proteins involved in TLR signaling different from that of the MyD88-dependentp athway.T hus, we tested the effect of INH14 in the endosomal receptor TLR3 pathway.T LR3 stimulation by polyI:C leads to the activation of the three transcription factors NF-kB, AP-1, and IRF3. We used HEK293c ells to test the activation of TLR3 because they expressed this receptor,e ven thougha tl ow levels. [22] HEK293 cells were transfected with reporter plasmids that encoded promoters with binding sites for the above-mentioned transcription factors (i.e., kb3.luc, IN-Fb.luc,a nd AP1.luc). Then we assessed the effect of INH14 on luciferase activity after stimulation with PolyI:C ( Figure 4C). INH14 decreased the transcriptionala ctivity of NF-kB, butn ot of IRF3. Nevertheless, we did not obtain ad etectable signal with the PolyI:C stimulation of cells transfected with the AP-1.luc plasmid( not shown). Previous studies have shown that, after TLR3 ligation, the signaling axis TRIF-RIP1-TAK1/TAB1 leads to IKKa/b activation. [23] Thus, we transfected ap lasmid encoding the TRIF adaptor in HEK293 cells, which produced a similar result after incubation with INH14( Figure 4C). These data indicate ad ownstream target of INH14 in the TLR3 pathway also present in the previously identified pathways (i.e., TLR2, TLR4, IL-1R, TNF-R), most likely IKKa/b.

IKKa and IKKb as targetsofI NH14
Our previouse xperiments in cellulo indicatedahigh degree of certainty that IKKs were the cellular targets of INH14. Both kinases have been described to be involved in different cellular functions:I KKa is essentiali nt he noncanonical NF-kB pathway, the deregulationo fw hich is associatedw ith lymphoid malignancies,a nd IKKb is active in the NF-kB canonical pathway in controlo fi mmune responses [24] and additional functions such as angiogenesis or insulin resistance. [25] To investigate the kinase-inhibitory activity of INH14, we performed kinase assays with recombinantI KKa.P reincubation of the enzyme with in-creasingc oncentrationso fI NH14 led to ar eduction of phosphorylated product (IC 50 = 8.97 mm;F igure 5A). Because several of the described IKKa inhibitors also inhibitedI KKb,w ea lso performed kinase assays with this enzyme. As seen in Figure 5B,I NH14 inhibited IKKb catalytic activity in ad ose-dependent way,w ith an even lower IC 50 than that obtained for IKKa (IC 50 = 3.59 mm). Ta ken together, these results indicatet hat INH14 is ap otent inhibitor of IKKa/b,a nd therefore, an inhibitor of the canonicaland noncanonical NF-kB pathways.However,u nrelated kinases, which have not been tested in this study, might also be targeted, even if with lower affinity.

Reduced lipopeptide-induced inflammation in miceby INH14
Intraperitoneal lipopeptide injection in mice leads to am aximum TNFa production2 ha fter injection. [26] To study the effect of INH14 on systemic inflammation, C57BL/6J mice were intraperitoneally injected with INH14 (5 mgg À1 ), followed by i.p. injectiono fP 2( 2.5 mgg À1 ). TheT NFa in the serum of the mice was quantified by meanso fE LISA( Figure 5C). Mice treated with INH14 had ad ecreased level of TNFa,i nc omparison with the control group (from (231.1 AE 21.3) to (115.8 AE 30.61) pg mL À1 ). Thus, the inhibition of IKKa/b by INH14 in vivo leads to ad ecrease in TNFa production after lipopeptide injection. During these experiments, we did not observe pathological effects in animals treated with the compound (e.g.,w eight loss, abnormal movements,dyspnoea).
INH14 did not affect the activation of the INF-b promoter after TLR3 activation ( Figure 4C). Therefore, gene transcription induced by TLR3 activation through IRF3 was unaltered. Thus, although INH14 has in vivo anti-inflammatory activity (e.g.,p otentially in Gram-positive bacterial sepsis with increased TLR2 functioning), antiviral activity might not be compromisedb y INH14 treatment. Nevertheless,I KK inhibitors have to be carefully characterized in vivo because systemick inase inhibition might potentially lead to septic shock. [27] Further in vivo studies will follow to assess the effect of INH14 in another kind of inflammation model (e.g.,L PS-induced shock, inflamed paw model).

INH14 is not toxic to primary human monocytes, but inhibitsthe growth of ovarian cancer cells
We next investigated if INH14 wast oxic for human primary immune cells. Monocytes from healthyv olunteers were prepared ands eeded in 96-well plates. Then, the cells were treated overnight with INH14 or with vehicle. The next day,c ell viability was assessedt hrough the CCK-8 assay (dehydrogenase activity in viable cells), which showed that INH14( 20 mm)w as not toxic to the cells, in comparison to vehicle incubation (Figure 6A).
The important role of IKKs in cancerr egulation is backed up by multiple studies. [28] IKKs regulate NF-kB activation, which, in turn, controls crucial steps in tumor development, such as transformation, survival, proliferation, and metastasis. Moreover,s ilencing different IKK subunitso rt heir pharmacological inhibition promote cell death and sensitize cancerc ells to chemotherapeutic agents. [29] Thus, we used the ovarian cancerc ell line SKOV3, in which NF-kB signaling has been shownt ob e upregulated, [30] to test the effect of INH14 on the cell-migration ability (wound-healing assay). The cells were grown overnight beforeasteady scratch was performed in each well. The cells were then incubated with vehicle;I NH14;o r5 Z-7-oxozeaenol as ac ontrol (TAK1 inhibitor [31] ), which has been demonstrated to inhibitt he migration potential. After 48 h, migration of the cells was observed by means of light microscopy.A ss hown in Figure 6B,t he woundc losing of SKOV3 decreased for cells incubated with INH14, in comparison to those treated with the vehicle. However,t he effect was lower than that observed with oxozeaenol at the same concentration.
NF-kB signaling is constitutively activated in av ariety of tumor cells and is associated with poor clinicalo utcome. [32] Therefore, we wanted to investigate if INH14 could decrease this basal NF-kB activation in SKOV3 cells. Cells seededi n9 6well plates were transfected with Elam.luc and Renilla plasmids and, after overnight incubation with INH14, we measured NF-kB activity.A so bserved in Figure 6C,I NH14d ecreasedt he basal luciferase activity by 50 %. Thisr esult mirroredt he results obtained by immunoblotting of SKOV3 cells incubated with INH14 ( Figure 6D). Treatmentw ith the compound decreased IkBa degradation similarly to that with oxozeaenol.
In future studies, the effect of INH14i nc ombination with other chemotherapeutic agents, such as cisplatin, and its effect in other typeso fh ematological and solid malignancies, in which the axis MyD88-NF-kB is constitutively activated, [33] will be assessed.

Conclusion
Our data indicatet hat the mechanism by which INH14 attenuates TLR2/TLR4/TNF-R/IL-1R signaling is through inhibition of IKKa/b.D ocking studies strongly suggested the binding of the biaryl urea scaffold to the hinge region of the kinase. INH14 decreased in vivo TLR2-induced inflammation, and in the future the inhibitory activity of INH14 in other animal models of inflammationw ill be studied. Our studies demonstrate that INH14 is ap romising startingp oint for the development of anti-inflammatory drugs.

Experimental Section
Cell culture:H EK293 cells, RAW264.7 mouse macrophages, bonemarrow-derived macrophages, and the ovarian cancer cell line SKOV3 (ATCC collection and Sigma-Aldrich) were cultured in Dulbecco's modified Eagle's medium (DMEM;S igma-Aldrich) with 10 %f etal bovine serum (FBS;S igma-Aldrich) and 0.5 %( v/v)c iprofloxacin (Sigma-Aldrich) at 37 8Ca nd 5% CO 2 atmosphere. The cell line HEK293-TLR2 (kindly provided by D. T. Golenbock;W orcester, MA, USA) was cultured as described above, with the addition of puromycin (10 mgmL À1 ;S igma-Aldrich). Human PBMCs were isolated from whole blood of healthy donors after volunteers' informed consent. For the isolation of monocytes, blood was layered on Histopaque 1077 (Sigma-Aldrich) at a1 :1 ratio. The mix was centrifuged at 400 g for 30 min. Peripheral blood mononuclear cells (PBMCs) were collected and washed twice with phosphate-buffered saline (PBS) and suspended in RPMI-1640 media with 3% FBS and 0.5 %c iprofloxacin (Sigma-Aldrich). Then, the cells were seeded in 96-well plates at ad ensity of 80 10 4 cells per well and the next day the media was changed before treatment with different stimulants.
Wound-healing assay:S KOV3 cells were seeded in 6-well plate dishes at ad ensity of 8 10 5 cells per well. Then, the cell monolayer was scratched with ap ipette tip (200 mL) to produce an arrow wound-like slit. The growth medium was replaced with DMEM plus 2% serum, and photographs were taken after 48 hw ith a4 0 objective in aL eica DMI4000B microscope. The wound gap was measured with the ImageJ software. In each image, the wound spacing was measured in the upper,l ower,a nd middle parts of the wound and the mean value was calculated. For each condition, four independent experiments were performed.
Kinase assays:I KKa and IKKb kinase assays (ADP-Glo kinase assay) were purchased from Promega and used by following the manufacturer's instructions. Quantification of adenosine diphosphate (ADP) produced in the reactions (chemiluminescence) was measured with aV ictor plate reader (PerkinElmer). The assay conditions are described in the legend of Figure 5.
Immunoblotting:H EK293-TLR2 or SKOV3 cells (10 5 cells per well) were grown in DMEM plus 10 %F BS overnight in 24-well plates. The next day,t he cells were treated as indicated. After stimulation, the cells were washed with PBS and then lysed with lysis buffer (150 mm NaCl, 0.1 %T ween, 20 mm Tris·HCl, pH 7.5) and ap rotease and phosphatase inhibitor mix (Roche Applied Science). The lysates were cleared by centrifugation at 10 4 g for 10 min (4 8C). Equal amounts of the supernatant were separated by electrophoresis on SDS 10 %p olyacrylamide gels and transferred to ap olyvinylidene difluoride (PVDF) membrane (CarlRoth). The membrane was blocked for 1h in 5% nonfat milk in TBST (150 mm NaCl, 0.1 % Tween, 20 mm Tris·HCl pH 7.5). Then it was incubated overnight at 4 8Cw ith the corresponding primary antibody and subsequently with HRP-conjugated secondary antibody for 2h.I mmunoreactive proteins were detected by using Immobilon detection reagents (Millipore) and the Fusion analyzer imager (Vilber). Quantification was achieved with the FusionCapt software (Vilber).
Mice experiments:T he 8-week-old, male, pathogen-free C57BL/6J mice (Charles River Laboratories) were maintained at the animal facility of the Medical University Innsbruck (12 hl ight/dark cycle; standard rodent chow and water available ad libitum). For lipopeptide-induced inflammation, 5 mgg À1 of INH14 or vehicle was administered intraperitoneally.A fter 1h,P 2( 2.5 mgg À1 )w as injected, and tail vein blood (25 mL) was collected at that time point (0 h) and 2hafter.T he blood was centrifuged at 5000 g,a nd the supernatant was frozen at À20 8Cu ntil further cytokine measurement by means of ELISA. Animal experiments were conducted according to national guidelines and European Community laws and were approved by the Committee for Animal Protection of the Austrian Ministry of Science.
Statistical analysis:G raphPad Prism (San Diego, CA, USA) was used to perform statistical analysis. Significance in the differences between data groups was assigned by using the Student t-test. Docking studies:D ocking was performed with as tructure of IKKb originally bound with the staurosporine analogue K252a (PDB ID: 4KIK). The structure was prepared with the Structure Preparation wizard of Maestro. [36] Following structure preprocessing, optimized hydrogen bonds were automatically assigned, any water molecules that formed less than three hydrogen bonds with non-water molecules were removed, and the receptor structure was subjected to restrained minimization (default settings applied for all of these procedures were executed with the Structure Preparation wizard). The receptor grid for docking was generated for chain A, with the binding site defined by the location of the cocrystallized ligand (default settings). Docking was performed with Glide SP [37] within Maestro (with default settings).