structure‐based design and synthesis of a small molecule

Structure-Based Design and Synthesis of a SmallMolecule that Exhibits Anti-inflammatory Activity byInhibition of MyD88-mediated Signaling to BacterialToxin Exposure

Shahabuddin Alam1, Sacha Javor2, MelissaDegardin2, Dariush Ajami2, Mitra Rebek2,Teri L. Kissner1, David M. Waag1,Julius Rebek Jr2 and Kamal U. Saikh1*

1Department of Immunology, Army Medical ResearchInstitute of Infectious Diseases, 1425 Porter Street,Frederick, MD 21702, USA2Department of Chemistry, The Skaggs Institute forChemical Biology, The Scripps Research Institute, 10550North Torrey Pines Road, La Jolla, CA 92037, USA*Corresponding author: Kamal U. Saikh,[email protected]

Both Gram-positive and Gram-negative pathogens orpathogen-derived components, such as staphylococ-cal enterotoxins (SEs) and endotoxin (LPS) exposure,activate MyD88-mediated pro-inflammatory cellularimmunity for host defense. However, dysregulatedMyD88-mediated signaling triggers exaggeratedimmune response that often leads to toxic shock anddeath. Previously, we reported a small molecule com-pound 1 mimicking BB-loop structure of MyD88 wascapable of inhibiting pro-inflammatory response toSEB exposure in mice. In this study, we designed adimeric structure compound 4210 covalently linkedwith compound 1 by a non-polar cyclohexane linkerwhich strongly inhibited the production of pro-inflam-matory cytokines in human primary cells to SEB (IC50

1–50 lM) or LPS extracted from Francisella tularensis,Escherichia coli, or Burkholderia mallei (IC50

10–200 lM). Consistent with cytokine inhibition, in aligand-induced cell-based reporter assay, compound4210 inhibited Burkholderia mallei or LPS-inducedMyD88-mediated NF-kB-dependent expression ofreporter activity (IC50 10–30 lM). Furthermore, resultsfrom a newly expressed MyD88 revealed that 4210inhibited MyD88 dimer formation which is critical forpro-inflammatory signaling. Importantly, a singleadministration of compound 4210 in mice showedcomplete protection from lethal toxin challenge. Col-lectively, these results demonstrated that compound4210 inhibits toxin-induced inflated pro-inflammatoryimmune signaling, thus displays a potential bacterialtoxin therapeutic.

Key words: drug design, structure based, inhibitors, toxicshock, MyD88, inflammation, staphylococcal enterotoxin B,cytokine

Received 3 July 2014, revised 27 October 2014 andaccepted for publication 1 November 2014

Myeloid differentiation primary response protein 88(MyD88) is predominantly involved in regulating intracellularimmune signaling that is critical for several aspects ofinnate immunity and host defense. It functions as anessential anchor protein that integrates and transducesintracellular signals generated by the Toll-like or IL-1receptor (TLR or IL-1R) super family. After ligand bindingto these receptors, MyD88 is recruited to the membrane–receptor complexes as a dimer by interaction of itsC-terminal Toll-interleukin receptor (TIR) domain with ananalogous domain in the IL-1R or TLR receptors (1). TheN-terminal death domain of MyD88 recruits a deathdomain-containing protein known as IL-1R-associated kin-ases (IRAKs) (2). Activation of IRAK leads to a series ofdownstream signaling cascades that activate NF-kB, p38mitogen-activated protein kinase (MAPK), and other regu-lators of gene expression that induce inflammatory cyto-kines, for example TNF-a, IL-1, and IL-6. While TIR–TIRdomain interaction is important for MyD88-mediated pro-inflammatory signaling, it has been reported that MyD88also interacts with interferon gamma receptor 1 (IFNGR1),which lacks a TIR domain (3). Similar interactions havebeen detected between MyD88 and other proteins lackingTIR and death domains such as Bruton’s tyrosine kinase(4), phosphatidyl-inositol-3-OH kinase (5), and interferon(IFN) regulatory factor 7 (6–8). These reports indicate thatMyD88 can associate directly or indirectly with other sig-naling proteins that are involved in pro-inflammatory signal-ing. Consistent with these results, a recent reportindicated that deficiency in MHC class II resulted inimpaired TLR-triggered production of pro-inflammatorycytokines and protected mice from an otherwise lethalchallenge with TLR ligands and live Gram-negative bacte-ria (9). We also reported that the binding of SEA or SEB toMHC class II activates MyD88-mediated pro-inflammatory

200 ª 2014 John Wiley & Sons A/S. doi: 10.1111/cbdd.12477

Chem Biol Drug Des 2015; 86: 200–209

Research Article

cytokine signaling in human primary cells (10). Resultsfrom our laboratory demonstrated that MyD88 geneknockout (MyD88-/-) mice showed a reduction in serumlevels of pro-inflammatory cytokines (11,12) and wereresistant to lethal SEA or SEB challenge. These resultssuggest that the TLR- and MHC-mediated responsesengage MyD88 (13).

Activation of MyD88-mediated pro-inflammatory signalingpathway is very important for several aspects of hostdefense; therefore, mice deficient in MyD88 have pro-foundly impaired innate immune responses and are sus-ceptible to a wide range of infectious diseases (14).However, while inflammatory innate immunity is importantfor host defense, there is a fine line between the beneficialand harmful effects of inflammation. Therefore, the innateimmune response to MyD88 should be elicited in a bal-ancing way. Otherwise, dysregulation of the MyD88-medi-ated immune signaling such as excessive or anuncontrolled signaling can trigger severe inflammatory dis-orders that can lead to profound clinical syndrome includ-ing severe pathological consequences such as septicshock. As MyD88 is involved in these disorders, it alsorepresents a valid target for therapeutic intervention tomodify the immune responses. Earlier results from ourlaboratory demonstrated that a synthetic mimetic of theBB-loop in the TIR domain of MyD88 (compound 1) atten-uated SEB-induced pro-inflammatory cytokine productionin human primary cells and increased survivability of micefrom toxic shock-induced death after a lethal SEB chal-lenge (15). Seeking to improve the efficacy of compound1, as a broad-spectrum bacterial toxin therapeutic capableof inhibiting severe toxicity to both enterotoxin and endo-toxin exposure, we synthesized a dimeric structure inwhich two modules of compound 1 were covalently linkedby a non-polar cyclohexane spacer. This molecule, com-pound 4210, was tested in primary cultures of humanperipheral blood mononuclear cells (PBMCs) for inhibitionof cytokine release associated with exposure to SEB, LPSfrom E. coli, Francisella tularensis, or Burkholderia mallei.Our results provide evidence that by targeting MyD88,compound 4210 inhibited inflammatory cytokine produc-tion in human primary cells and increased therapeuticefficacy in mice with exposure to toxins of both Gram-positive and Gram-negative pathogens.

Materials and Methods

ReagentsStaphylococcal enterotoxin B (SEB) and SEA were pur-chased from Porton Down, Inc. (Salisbury, UK) and storedat �70 °C. SEB was endotoxin-free and prepared underGMP conditions. Escherichia coli LPS (055:B5) was pur-chased from Sigma-Aldrich (Saint Louis, MO, USA). Pooledhuman AB sera were obtained from Pel-Freez (Brown Deer,WI, USA). LPS extracts purified from F. tularensis andB. mallei used in this study were obtained from Dr. Dave

Waag (USAMRIID) as described elsewhere (16). A cytomet-ric bead array (CBA) kit was purchased from Pharmingen(San Diego, CA, USA). The Meso Scale Discovery (MSD)multi spot array ultra-sensitive cytokine assay kit was pur-chased from MSD (Gaithersburg, MD, USA). Ficoll–Hyp-aque was purchased from GE Healthcare Biosciences(Piscataway, NJ). Primary anti-MyD88 antibody wasobtained from AnaSpec, Inc., (San Jose, CA) and AlexisBiochemicals (San Diego, CA, USA). Anti-b-actin antibodywas purchased from Cell Signaling Technology (Danvers,MA, USA). Plasmid 12287 (pCMV-HA-MyD88) and plasmid13093 (MyD88-flag) were purchased through an MTAagreement with Addgene (Cambridge, MA, USA). HEK 293-SEAP-TLR4 cell was obtained from Invivogen (San Diego,CA, USA). The transfection reagent lipofectamine 2000 waspurchased from Invitrogen (Carlsbad, CA, USA). All chemi-cal reagents were purchased from Aldrich and used withoutfurther purification. The solvents were purchased fromFisher or J.T. Baker. All deuterated solvents were pur-chased from Cambridge Isotope Laboratories, Inc.

Synthesis of compound 4210Trans-cyclohexane-1, 4-dicarboxylic acid (0.52 g,3.0 mmol) was stirred in neat thionyl chloride (30 mL) at90 °C for 6 h. The solvent was evaporated, and theresidue was placed under a nitrogen atmosphere. (S)-3-methyl-2-((3-phenylpropyl)amino) -1-(pyrrolidin-1-yl)butan-1-one (17) (1.7 g, 6.0 mmol) in dry tetrahydrofurane(30 mL) and diisopropylethylamine (2.1 mL, 1.6 g,12 mmol) were added, and the solution was stirred atroom temperature for 24 h. The reaction mixture wasdiluted with ethyl acetate (0.15 L), extracted with NaOH1.0 M (50 mL), HCl 1.2 M (50 mL), and sat. NaHCO3

(50 mL), and dried over Na2SO4. After evaporation of thesolvents, the residue was purified by silica gel flash chro-matography (acetone/CH2Cl2: 3?6%) affording compound4210 (1.44 g, 67.0%) as a white foam. 1H NMR(600 MHz, CDCl3) d 7.35 – 7.25 (t, J = 7.4 Hz, 4H), 7.22– 7.09 (m, 6H), 5.14 – 4.95 (d, J = 10.7 Hz, 2H), 3.56 –3.46 (m, 4H), 3.42 – 3.31 (m, 4H), 3.30 – 3.17 (m, 2H),2.68 – 2.58 (m, 2H), 2.52 – 2.42 (m, 2H), 2.38 – 2.27 (m,2H), 2.27 – 2.17 (m, 2H), 1.95 – 1.87 (m, 2H), 1.86 – 1.77(m, 6H), 1.76 – 1.64 (m, 4H), 1.63 – 1.52 (m, 6H), 1.32 –1.22 (m, 2H), 0.98 – 0.87 (d, J = 6.5 Hz, 6H), 0.80 – 0.68(d, J = 6.8 Hz, 6H); 13C NMR (150 MHz, CDCl3) d 176.37,168.93, 140.70, 128.70, 128.33, 126.31, 77.16, 59.22,46.70, 46.11, 42.84, 40.34, 33.18, 32.42, 29.12, 28.19,27.40, 26.00, 24.10, 19.82, 18.02; Rf (acetone/CH2Cl2,9%) 0.59; HRMS (ESI+) Calcd. for [C44H64N4O4+H

+]713.5000, found 713.5016.

Elements of the supporting information for 4210synthesis1H NMR spectra were recorded at 600 Mhz and 13C NMRspectra at 150 MHz a Bruker Avance DRX-600 spectrom-eter equipped with a 5 mm QNP probe at 300 K.

Chem Biol Drug Des 2015; 86: 200–209 201

Anti-inflammatory Activity of a MyD88 Inhibitor

Chemical shifts of 1 H NMR and 13C NMR are given inppm (d) with respect to trimethylsilane (TMS, d = 0.00)using residual CHCl3 proton or carbon signal as references(7.26 ppm, 1H spectra, and 77.16 ppm for 13C spectra).Coupling constants (J) are reported in hertz (Hz). Standardabbreviations indicating multiplicity were used as follows: s(singlet), d (doublet), t (triplet), and m (multiplet). HRMS-ESI-TOF spectra were recorded by the mass spectroscopyservice of the Scripps Research Institute using an AgilentESI-TOF spectrometer.

MicePathogen-free, 12–16-week-old BALB/c and C57BL/6mice were obtained from Charles River (NCI-Frederick,Frederick, MD, USA). MyD88�/� was obtained fromCharles River Laboratory as described elsewhere (11,12).

Cell isolation and purificationPeripheral blood mononuclear cells were obtained fromconsenting healthy donors in accordance with an Institu-tional Review Board-approved research donor protocol.Peripheral blood mononuclear cells were isolated by stan-dard density gradient centrifugation with Ficoll–Hypaque,harvested from the interface, washed, and re-suspendedin RPMI 1640 medium. Isolated cell populations were>98% purity.

Cytokine analysisCell cultures were incubated (37 °C, 5% CO2) for 16 h.Cytokines in culture supernatants were measured by aCBA kit using captured beads coated with antibodies spe-cific for cytokines and flow cytometry analysis asdescribed elsewhere (15). Cytokine measurements wereconfirmed by dilution of culture supernatant using HumanInflammation and Th1/Th2 CBA kits and acquiring 1800beads. We also used Meso Scale Discovery (MSD) multispot array ultrasensitive cytokine assay kit for measuringcytokines in culture supernatants (according to the manu-facturer’s protocol) as described (15).

Western blot analysisThe transfected cells were chilled on ice for 5 min beforebeing pelleted into fresh 1.5 mL centrifuge tubes. Mem-brane and cytoplasm separation was performed by sus-pending the pellets in 50 lL of lysis buffer (active motif)in the presence of DTT, protease inhibitors, and phos-phatase inhibitors and incubated on ice for 30–60 min.The membrane fraction was collected by centrifuging thelysates at 14 000 9 g for 20 min. The supernatant con-tained the cytoplasmic fraction. Samples containing10 lg of total cytoplasmic proteins were separated bygel electrophoresis and transferred to nitrocellulose mem-branes. The membranes were blocked overnight in 19Tris-buffered saline (TBS) containing 0.1% Tween-20

(TBST) and 3% bovine serum albumin at 4 °C. Themembranes were washed extensively with 19 TBST buf-fer and then probed with anti-MyD88 polyclonal antibodyfollowed by horseradish peroxidase-conjugated secondaryantibody (goat anti-rabbit). After additional rinsing with1 9 TBST buffer, the membranes were exposed to achemiluminescent substrate in the presence of hydrogenperoxide, using Immun-Star WesternC Chemilumines-cence kit (Bio-Rad, Hercules, CA, USA). A VersaDocModel 4000 (Bio-Rad) imaging system was used to cap-ture the image.

Cell culture and transfectionsHuman embryonic kidney (HEK) 293-SEAP-TLR4 cellswere cultured in EMEM, supplemented with 10% fetalbovine serum (FBS) (Invitrogen) and grown in a 37 °Chumidified atmosphere of 5% CO2. For co-immunoprecipi-tation of MyD88-Flag/HA-MyD88, HEK 293-SEAP-TLR4cells were cultured in 6-well plates and transfected by lipo-fectamine 2000 (Invitrogen) method with 4 lg of theappropriate plasmids according to the manufacturer’sinstructions. The compound 4210 was added to the med-ium 6 h after transfection. Six hours after transfection, cellswere cultured with LPS in the presence or the absence ofcompound 4210.

Co-immunoprecipitation assayHEK293-SEAP-TLR4 cells (transfected or Mock) were col-lected 48 h after transfection, washed with 2 mL of ice-cold PBS, and lysed in 80 lL of buffer [50 mM HEPES,pH7.4]. Cells were pelleted by centrifugation at10 000 9 g for 10 min at 4 °C, and cytosolic fractionswere collected for immunoprecipitation. Cell extracts (1 mgtotal proteins) were incubated with 2 lg of mouse anti-FlagM2 conjugated with agarose attached to magnetic beads(Sigma–Aldrich) for 16 h under constant shaking at 4 °C.Agarose bead-bound immunocomplexes were separatedby a magnetic separator, washed three times, and elutedin SDS-PAGE sample buffer for Western blot analysis.

Secreted alkaline phosphatase assayTLR4/MD-2/NF-kB/SEAPorter HEK 293 cells (5 9 105

cells/mL/well) were cultured with LPS (1 lg/mL), with vary-ing concentrations of compound 4210 in a 96-well plateand incubated at 37 °C for 16 h. The culture supernatantwas collected and centrifuged to remove any cell debris.The Great EscAPe secreted alkaline phosphatase (SEAP)assay from Clonetech was used to determine the amount ofalkaline phosphatase secreted into the supernatant. A 19dilution buffer was prepared from a 59 stock solution and75 lL of the 19 dilution buffer was mixed with 25 lL of thesupernatant, incubated for 30 min at 65 °C to inactivateendogenous alkaline phosphatase. The samples wereplaced on ice for 3 min and then equilibrated at room tem-perature. Secreted alkaline phosphatase substrate solution

202 Chem Biol Drug Des 2015; 86: 200–209

Alam et al.

(100 lL) was added to each sample and read at 10 minintervals using a chemiluminescence reader.

Results

Design and Synthesis of compound 4210We aimed to increase the potency of the earlier reportedcompound 1 (15) by covalently linking as a dimer forgreater flexibility to interact with target domain of MyD88.For this, we covalently linked compound 1 by a non-polarcyclohexane ring aiming that a six-member heterocyclering will increase flexibility for binding to the exposed BB-loop of the TIR domain of MyD88 because MyD88 isrecruited as a dimer for signal transduction. The com-pound 4210 was synthesized from the previously reportedcompound 1 (15) as shown in materials and methods (Fig-ure 1). When assessed in biological assays and in vivo,dimeric compound 4210 with a non-polar linker in thespacer region between the monomer significantlyincreased the potency of the molecule as an inhibitor ofMyD88 signaling as described below.

Inhibition of pro-inflammatory cytokine productionof human peripheral blood mononuclear cells withexposure to SEB, LPS, or killed Burkholderiamallei in the presence of compound 4210To assess the biological effect of compound 4210 on theinhibition of pro-inflammatory cytokine production of pri-mary human cells, PBMCs were cultured with SEB, orLPS, from F. tularensis or E. coli, or with whole B. mallei

(irradiated) in the presence of compound 4210. The cul-tures were incubated for 16 h and culture supernatantswere collected for measuring cytokines. The release ofpro-inflammatory cytokines such as IL-1b, IL-6, IL-2,TNF-a, and INF-c in culture supernatant was inhibited in adose-dependent manner in the presence of the compound4210 (Figure S1). The average inhibitory concentration(IC50) of compound 4210 appeared to be 1–50 lM formost of the pro-inflammatory cytokines tested with somevariation in response from donor to donor (Figure 2A). Thecompound 4210 was upgraded from compound 1 (15). Toprove that the anti-inflammatory effect of compound 4210was superior that of compound 1, we tested for cytokineinhibitory effect by direct comparison of compound 1 andcompound 4210 under the same conditions from PBMCs

of two additional normal donors stimulated with SEB. Indonor 1, the IC50 of compound 4210 for IFN-c, IL-1b,IL-6, and TNF-a was 0.30, 0.37, 0.4, and 0.37 lM,respectively, whereas IC50 of compound 1 for IFN-c, IL-1b,IL-6, and TNF-a was 41, 35, 14, and 19 lM, respectively(Figure S2). Similar results were observed with donor 2(Figure S2). These results demonstrate that the compound4210 is indeed an enhanced version of compound 1.Besides SEB, the compound 4210 also inhibited cytokineproduction when primary PBMCs cultured with LPS (F. tu-

larensis or E. coli), or killed B. mallei (Figure 2B, FigureS3). These results suggest that compound 4210 wascapable of inhibiting SEB, B. mallei, or LPS (F. tularensisor E. coli)-induced pro-inflammatory cytokine production.These ex vivo cytokine inhibition data are consistent withour earlier reports which demonstrated MyD88-mediatedpro-inflammatory cytokine response upon SEB or SEAexposure in mice, and cytokine production were inhibitedin MyD88�/� mice (10–12).

Inhibition of MyD88-specific signaling bycompound 4210To confirm the inhibitory effect of compound 4210 onMyD88 signaling, we utilized a cell-based reporter assay.A stable cotransfected (TLR4-MD2-NF-kB/SEAPorterTM)HEK 293 cell line was used to detect a MyD88-mediatedNF-kB-driven SEAP reporter response induced by LPS(TLR4-ligand). Compound 4210 inhibited the LPS (E. coli,F. tularensis, or B. mallei)-induced MyD88-mediated (viaTLR4-MD2-MyD88-NF-kB-SEAP) SEAP response in adose-dependent manner (Figure S4). The inhibitory con-centration (IC50) of compound 4210 was determined fromthe dose-dependent inhibition curve (Figure 3). It is impor-tant to note that IC50 of SEAP reporter activity and cyto-kine inhibition was comparable (Figures 2B and 3).Further, to prove that the enhanced effect and mechanismof compound 4210 upgraded from compound 1, wetested a direct comparison of compound 1 and compound4210 under the same conditions in SEAP reporter assay.The compound 4210 demonstrated a dose-dependentinhibition of SEAP reporter expression at much lower con-centration than compound 1 (Figure S5). The calculatedIC50 value of SEAP response of compound 4210 activatedwith LPS (E. coli, F. tularensis) and B. mallei was 1, 0.47,and 6 lM, respectively, compared to IC50 Compound 1

stimulated with LPS (E. coli, F. tularensis) and B. mallei

Figure 1: Synthesis of compound 4210.Compound 4210 was synthesized fromCompound 1 by covalent linkage vianonpolar cyclohexane as described insynthetic scheme in the section of Materialsand Methods.



was 87, 270, and 180 lM, respectively (Figure S5). Wealso tested the various batches of compound 4210 syn-thesized at different times for quality assurance and qualitycontrol (QA/QC) by SEAP reporter assay. The compound4210 showed a consistent dose-dependent inhibition ofMyD88-specific signaling (SEAP) reporter expression (datanot shown). These results suggest that compound 4210by targeting specifically MyD88 inhibited downstream NF-kB-driven SEAP reporter activity.

Inhibition of newly expressed MyD88–MyD88association in the presence of compound 4210Our earlier reports demonstrated that SEB stimulationupregulates MyD88 in primary human cells (10–12). Next,

we tested whether compound 4210 targets newly synthe-sized MyD88 when activated with ligand such as LPS. Tofollow this, HEK 293-SEAP-TLR4 cells were cotransfectedwith plasmids pCMV-HA-MyD88 and MyD88-Flag, and6 h later, cells were treated with varying concentrations ofcompound 4210. In a co-immunoprecipitation assay usinganti-Flag antibody, followed by SDS-PAGE and immuno-blot analysis with anti-MyD88 antibody, an increase inaccumulation of the newly expressed MyD88 (31 kDa) pro-tein was observed in a dose-dependent manner (Fig-ure 4A) compared to untreated cells. Results shown inFigure 4A suggest that in the presence of compound4210, newly expressed MyD88 dimer formation was inhib-ited and maximum inhibition was observed in between 10and 100 lM. The 31 kDa MyD88 band was further con-firmed by reprobing with anti-HA antibody, and the resultsshowed a compound 4210 dose-dependent results (Fig-ure 4B). These data demonstrate that Flag-MyD88 andHA-MyD88 association was inhibited by the compound4210. Interestingly, these data are consistent with cytokineinhibition and SEAP reporter expression (Figures 2 and 3).In agreement with earlier reports, these results indicatethat ligand binding to TLR4 induced higher level of MyD88expression, which allowed it to associate through its TIRdomain for homodimer formation, whereas in the presenceof compound 4210 by targeting TIR domain, it inhibiteddimerization of MyD88. It is likely the possible mode ofbinding of the dimeric compound 4210 to MyD88 as 1:2molar ratio (4210:MyD88), as MyD88 is recruited as dimerfor signal transduction and thereby increased inhibition effi-ciency for the downstream signal transduction for pro-inflammatory cytokine responses.

A

B

Figure 2: Inhibition of pro-inflammatory cytokines with exposureto SEB, LPS, or Burkholderia mallei in primary cultures bycompound 4210. Peripheral blood mononuclear cells (PBMCs)(1 9 106) from three normal donors were cultured for 20 h withSEB (200 ng/mL) and with or without compound 4210 (250–0.5 lM). The culture supernatants were collected and measuredfor cytokine by cytometric bead array assay. PBMCs (1 9 106)from a normal donor were stimulated with LPS extracted fromFrancisella tularensis (10 lg), Escherichia coli (1 lg) or irradiatedB. mallei (10 lg) in the absence or the presence of variousconcentrations of 4210 by MSD assay as described in Materialsand Methods. The data presented as IC50 calculated as theconcentration required for inhibition of cytokine production inMNCs by 50% relative to the control. (A) IC50 to SEB; (B) IC50 toLPS extracted from F. tularensis or E. coli or irradiated B. mallei.Data presented representative of three independent experiments.

Figure 3: MyD88-specific signaling inhibited by compound 4210.Compound 4210 was tested for MyD88-specific signaling bymonitoring LPS-induced secreted alkaline phosphatase (SEAP)activity by a NF-kB-driven signaling pathway for reporter geneexpression. HEK 293 stable transfected cell line (TLR4-MD2-NF-kB-SEAP) was stimulated with LPS (TLR-4 ligand) and treatedwith varying concentrations of compounds 4210(500–0.01 lM).Culture supernatants were tested for SEAP activity and comparedto levels in the absence of compounds (positive control). The datapresented as IC50, calculated as the concentration required forinhibition of cytokine production in MNCs by 50% relative to thecontrol.


Alam et al.

Cytokine inhibition by compound 4210 withexposure to SEB, LPS, or Burkholderia mallei inmouse spleen cell culturesOur earlier results showed that compound 4210 stronglyinhibited cytokine inhibition in human primary cells withexposure to SEB, LPS, or B. mallei (Figure 2), andtargeted to newly synthesized MyD88 induced by LPS(Figures 3 and 4). As MyD88 is involved in pro-inflamma-tory cytokine production with exposure to Gram-negativepathogens or LPS, we next examined the effect of com-pound 4210 on cytokine inhibition in spleen cells of wild-type C57BL/6 and MyD88�/� mice (C57BL/6 background)

with exposure to LPS and B. mallei. As expected, cytokineproduction was abrogated in spleen cell cultures isolatedfrom MyD88�/� mice (Figure 5A), while wild-type C57BL/6produced pro-inflammatory cytokines. These results wereconsistent with our earlier report in which we describedabrogation of pro-inflammatory cytokine production inMyD88�/� mice with exposure to SEB (12). Next, weexamined whether treatment of compound 4210 wouldinhibit cytokine production in spleen cells of wild-typeBALB/c as well as C57BL/6 mice. Results shown in Fig-ure 5B,C demonstrated that compound 4210 inhibitedmost of the pro-inflammatory cytokine production withexposure to SEB, E. coli LPS, or B. mallei in spleen cellcultures. It is important to note that IC50 values varied withdifferent cytokines which is not unusual, because cytokineresponse varies with specificity of the cell types and kinet-ics. These results suggest that compound 4210 by specifi-cally targeting MyD88 inhibited the cytokine production inspleen cell cultures of both BALB/C and C57BL/6 miceexposed to both Gram-positive and Gram-negative patho-gens or pathogen-derived components such as B. mallei,LPS, or SEB. Thus, similar to human primary PBMC cul-ture, compound 4210 by targeting specifically MyD88inhibited cytokine signaling in spleen cells of mice.

In vivo efficacy of compound 4210: Attenuation ofSEB- or LPS-induced toxicity in miceNext, we examined to ascertain wether the targeting ofcompound 4210 with TIR domain of MyD88 is inhibitingcytokine production as we observed in ex vivo culture inmouse spleen cells and whether it could protect animalsfrom toxic shock-induced death with exposure to SEB orLPS. To address this, we used the LPS potentiation (sublethal dose) model of SEB toxicity in mice as reported ear-lier (11,12,18). BALB/c mice (n = 6) were injected with dif-ferent amounts of compound 4210 (0.17, 0.51 or 1 mg),and 30 min later, they were treated with SEB (1 lg) equiv-alent to (2 LD50), followed by LPS another 2 h later. Allanimals in the study, not treated, died by 44 h (Figure 6).In contrast, mice treated with compound 4210 showeddose-dependent protection. Mice that were treated withcompound 4210 at concentrations of 0.17 mg had 33%survivability and those that were treated with 0.51 mg had66% survivability, whereas mice that were treated withcompound 4210 at 1 mg were completely protected(p = 0.05) (Figure 6A). Postexposure treatment (first SEBexposure followed by compound 4210 treatment with 0.51mg or 1 mg and 2 h later LPS) delayed in death and 33%protection was observed at 1 mg (Figure 6B). Theseresults indicate that compound 4210 has in vivo therapeu-tic efficacy against SEB intoxication. We also examined in

vivo efficacy of compound 4210 against toxicity to lethaldose (120 lg) with only endotoxin (LPS) exposure. Ourresults showed that in comparison with lethal LPS-treatedmice, administration of a single dose of compound 4210(1 mg) delayed death and 33% of mice were protectedfrom lethal endotoxic shock (Figure 6C).

A

B

Figure 4: Inhibition of MyD88 homodimer formation from newlyexpressed MyD88 in the presence of compound 4210. HEK 293-SEAP-TLR4 cells were cotransfected with plasmids MyD88-Flagor pCMV-HA-MyD88. Six hours after transfection, cells wereincubated for 40 h with or without the compound 4210 (100 lM,10–0.1 lM). Cell extracts were immunoprecipitated (IP) with anti-Flag antibody, and immune-precipitated proteins were analyzed inWestern blot by probing with anti-MyD88 antibody. The blot wasstripped in stripping buffer and then reprobed with anti-HAantibody. Relative level of MyD88 was determined bydensitometry. (A) Immunoblot probed with anti-MyD88; (B)Immunoblot with probed with anti-HA antibody. Lower panelsrepresent densitometry analyses of the results shown in (A) and(B). The results presented as representative data of twoindependent experiments.



Discussion

Gram-positive and Gram- negative pathogens or patho-gen-derived component activates pro-inflammatoryresponse via MyD88-mediated signaling (1). Our previousstudies demonstrated that MyD88-mediated pro-inflam-matory cytokine signaling is critical to SEB toxicity(11,12). Here, we have shown that a synthetic moleculedesigned that resembles BB-loop-like structure of MyD88with dual-functional groups (dimer) connected with anon-polar cyclohexane linker significantly enhanced in

vivo efficacy (50 mg/kg) in attenuating SEB as well asLPS toxicity compared to compound 1 reported earlier(15). The compound 4210 exhibits anti-inflammatoryproperties effective in limiting pro-inflammatory cytokinesignaling when exposed to both enterotoxin and endo-toxin and thus has a potential for broad-spectrum thera-peutic use.

Toll-interleukin receptor domain constitutive association ofdimer formation is essential for MyD88-mediated signalingwith ligand binding to TLRs. Our earlier results and othershave shown that MHC class II plays a fundamental role indriving TLR-mediated inflammatory responses and MHCclass II negative cells do not upregulate MyD88 when stimu-lated with MHC class II ligands (9,10,13). Liu et al. (9)reported that deficiency in MHC class II impaired TLR-trig-gered production of inflammatory cytokines in macrophagesand DCs, and this protected mice from lethal challenge withTLR ligands and live Gram-negative bacteria. Furthermore,it has been shown that in lysates of HEK293 cells overexpressing HLA-DR, radio-labeled recombinant TLR-2 pro-tein precipitates in vitro together with HLA-DR protein immu-noprecipitated with anti-HLA-DR. So, TLR-2 was proposedto be associated with HLA-DR (19). However, whether ornot TLR-2 and HLA-DR interact physically has yet to bedetermined. Clinical results have shown that HLA-DR

A B C

Figure 5: Cytokine inhibition by compound4210 with exposure to SEB, LPS, orBurkholderia mallei in mouse spleen cellcultures. Spleen cells (2 9 106)/well fromBALB/c, C57BL/6, and MyD88�/� micewere cultured in a 48-well (300 lL/well) platefor 20 h with SEB (200 ng/mL), Escherichiacoli LPS (1 lg/mL), or B. mallei (10 lg/mL)with or without 4210 (100–0.1 lM) treatment30 min before antigen exposure as indicatedin the figures. The culture supernatants werecollected and measured for cytokine byMSD assay as described in Materials andMethods. (A) Cytokine response to B. mallei

or LPS in C57BL/6 (wild type) and MyD88�/

� mice, (B) cytokine inhibition in spleen cellcultures of BALB/c mice, and (C) C57BL/6mice in the presence of compound 4210.The data presented in (B) and (C) as IC50

concentration. Data presented arerepresentative of three independentexperiments.


Alam et al.

expression in peripheral blood monocytes is much lower inpatients with septic shock than in normal subjects (20,21).Such patients release much less TNF-a and IL-1b than donormal subjects in response to LPS. These results suggestthe involvement of TLR- and MHC-mediated responsesMyD88 for pro-inflammatory signaling (13) which likely is anoutcome in septic shock.

Our earlier results demonstrated that small molecules mim-icking BB-loop structure of MyD88 interfere with pro-inflammatory signaling with exposure to SEB, LPS, orSEB+LPS in ex vivo primary cells and in mice (16). Wealso reported that a dimeric compound EM-163 with twocompound 1 covalently linked with an aromatic ring (pyri-dine) linker had increased efficacy ex vivo in mice over thecompound 1 (16,22). EM-163 had a non-polar and flat(rigid) structural characteristic because of aromatic ringand less flexibility which likely requires a high concentrationfor a specific mode of binding in solution. In this study, ourresults showed that compound 4210 had an increase inefficacy both ex vivo (IC50 1–50 lM compared to EM-163IC50 50–400 lM) (22). Concurrent with ex vivo cytokine

inhibition data, administration of compound 4210 in miceincreased therapeutic efficacy (1 mg) compared to EM-163 (1.7 mg) from toxic shock-induced death with lethalSEB challenge (22). It may be that compound 4210 with anon-polar cyclohexane linker between the covalently linkedcompound 1 provided more flexibility and facilitates forbinding to exposed BB-loop residues of MyD88 andthereby prevent association of MyD88–MyD88 dimer for-mation. Although it has yet to be validated, it may be pos-sible that the binding of dimeric compound 4210 toMyD88 at 1:2 molar ratio (4210:MyD88) increased inhibi-tion efficiency and therapeutic efficacy. Besides SEB,compound 4210 appeared to be effective in inhibitingpro-inflammatory response to endotoxin exposure ofGram-negative pathogens, such as F. tularensis, E. coli, orB. mallei, in primary cultures of mice and humans. Thecombinations of endotoxins and enterotoxins exposurehave been shown to have severe consequences. Criticallyill patients may be exposed to endotoxin from Gram-nega-tive organisms and enterotoxin from toxin-producingGram-positive organisms, even if the organisms is simplycolonizing the patients. The dual exposure is intimately

A

B

C

Figure 6: In vivo efficacy of compound4210: Attenuation of SEB or LPS-inducedtoxic shock in mice. BALB/c mice (n = 6/pergroup) were injected (i. p.) with differentamounts of compound 4210 (0.17 mg,0.51 mg, or 1 mg), 30 min later injectedwith 1 lg of SEB, and followed by LPS(90 lg, sub lethal dose) 2 h later. Mice wereobserved to determine whether micesurvived, or time of death. Control miceinjected only with 90 lg of LPS or 1 lg ofSEB survived. Data represent three separateexperiments. (A) Survivability of mice withpretreatment of compound 4210; (B)survivability of mice with postexposure toSEB; and (C) survivability of mice withpretreatment of compound 4210 to lethalLPS (120 lg) exposure.



involved in the genesis of endotoxin–enterotoxin double hitwhich would lead to profound clinical consequences. Ourresults showed that compound 4210 was found to beequally effective in limiting pro-inflammatory response toboth endotoxin and enterotoxin exposure. It is likely that ina clinical scenario compound 4210 would be effective inrestricting amplification of the MyD88-mediated pro-inflam-matory signaling cascade. Our ongoing effort is to increasedrug-like properties of compound 4210 for potentialbroad-spectrum use.

Acknowledgments

This work was supported by The Defense Threat Reduc-tion Agency to K.U.S (grant CBM.THROX.03.10.RD.006).The authors thank Lorraine Farinick for her help for figurepreparation. We are also grateful to the investigators atAddgene for providing plasmids originally published inProc. Natl. Acad. Sci. 2006 Jul.18.103(29):10961-6.

Author Contributions

S.A. did cytokine measurement and did transfection, co-immunoprecipitation, and animal experiments and waspartly involved in writing the manuscript; S.J., M.D., D.A.,and M.R. contributed in the synthesis of compound 4210;T.L.K. performed cytokine measurement; D.M.W. contrib-uted in the extraction of LPS from F. tularensis andB. mallei culture and irradiation; J.R. designed and super-vised overall synthetic chemistry; and K.U.S. designed andsupervised the research and wrote the manuscript.

Disclosures

Research was conducted under an IACUC approved pro-tocol in compliance with the Animal Welfare Act and otherFederal statutes and regulations relating to animals andexperiments involving animals and adhere to principlesstated in the Guide for the Care and Use of Laboratory

Animals, National Research Council, 1996. The facilitywhere this research was conducted is fully accredited bythe Association for Assessment and Accreditation of Labo-ratory Animal Care International. Peripheral blood mononu-clear cells used in this study were obtained from healthydonors with written consents, in accordance with guide-lines of the human use committee (HUC) and institutional(USAMRIID) review board-approved research donor proto-col FY 05-05. Views expressed in this paper are those ofthe authors and do not purport to reflect official policy ofthe U.S. Government USAMRIID administrators.

Competing Financial Interests

The authors declare no competing financial interests.

References

1. Akira S., Takeda K. (2004) Toll-like receptor signaling.Nat Rev Immunol;4:499–511.

2. O’Neill L.A.J. (2003) The role of MyD88-like adapters inToll-like receptor signal transduction. Biochem SocTrans;31:643–647.

3. Sun D., Ding A. (2006) MyD88-mediated stabilization ofinterferon-gamma-induced cytokine and chemokinemRNA. Nat Immunol;7:375–381.

4. Takeda K., Kaisho T., Akira S. (2003) Toll-like recep-tors. Annu Rev Immunol;21:335–376.

5. Barto G.M., Medzhitov R. (2003) Toll-like receptor sig-naling pathways. Science;300:1524–1525.

6. Jefferies C.A., Doyle S., Brunner C., Dunne A., BrintE., Wietek C., Walch E., Wirth T., O’Neill L.A. (2003)Bruton’s tyrosine kinase is a Toll/interleukin-1 receptordomain-binding protein that participates in nuclear fac-tor kappaB activation by Toll-like receptor 4. J BiolChem;278:26258–26264.

7. Ojaniemi M., Glumoff V., Harju K., Liljeroos M., VuoriK., Hallman M. (2003) Phosphatidylinositol 3-kinase isinvolved in Toll-like receptor 4-mediated cytokineexpression in mouse macrophages. Eur J Immu-nol;33:597–605.

8. Honda K., Yanai H., Mizutani T., Negishi H., Shimada N.,Suzuki N., Ohba Y., Takaoka A., Yeh W.C., Taniguchi T.(2004) Role of a transductional-transcriptional processorcomplex involving MyD88 and IRF-7 in Toll-like receptorsignaling. Proc Natl Acad Sci U S A;101:15416–15421.

9. Liu X., Zhan Z., Li D., Xu L., Ma F., Zhang P., Yao H.,Cao X. (2011) Intracellular MHC class II molecules pro-mote TLR-triggered innate immune responses by main-taining activation of the kinase Btk. NatImmunol;12:416–424.

10. Kissner T.L., Ruthel G., Alam S., Ulrich R.G., Fernan-dez S., Saikh K.U. (2011) Activation of MyD88 signal-ing upon staphylococcal enterotoxin binding to MHCclass II molecules. PLoS ONE;6:e15985.

11. Kissner T.L., Cisney E.D., Ulrich R.G., Fernandez S.,Saikh K.U. (2010) Staphylococcal enterotoxin A induc-tion of pro-inflammatory cytokines and lethality in miceis primarily dependent on MyD88. Immunol-ogy;130:516–526.

12. Kissner T.L., Ruthel G., Cisney E.D., Ulrich R.G., Fer-nandez S., Saikh K.U. (2011) MyD88-dependent pro-inflammatory cytokine response contributes to lethaltoxicity of staphylococcal enterotoxin B in mice. InnateImmun;17:451–462.

13. Hassan G.S., Mourad W. (2011) An unexpected rolefor MHC class II. Nat Immunol;12:375–376.

14. Cook D.N., Pisetsky D.S., Schwartz D.A. (2004) Toll-like receptors in the pathogenesis of human diseases.Nat Immunol;5:975–979.

15. Kissner T.L., Moisan L., Mann E., Alam S., Ruthel G.,Ulrich R.G., Rebek M., Rebek J., Saikh K.U. (2011) Asmall molecule that mimics the BB-loop in the Toll/IL-1receptor domain of MyD88 attenuates staphylococcal


Alam et al.

enterotoxin B induced pro-inflammatory cytokine pro-duction and toxicity in mice. J Biol Chem;286:31385–31396.

16. Waag D.M., McKee K.T., Sandstrom G., Pratt L.L.K.,Bolt C.R., England M.J., Nelson G.O., Williams J.C.(1994) Cell-mediated and humoral immune responsesafter vaccination of human volunteers with the live vac-cine strain of Francisella tularensis. Clin Diagn LabImmunol;2:143–148.

17. Davis C.N., Mann E., Behrens M.M., Gaidarova S.,Rebek M., Rebek J. Jr, Bartfai T. (2006) MyD88-dependent and –independent signaling by IL-1 in neu-rons probed by bifunctional Toll/IL-1 receptor domain/BB-loop mimetics. Proc Natl Acad Sci USA;103:2953–2958.

18. Stiles B.G., Bavari S., Krakauer T., Ulrich R.G. (1993)Toxicity of staphylococcal enterotoxins potentiated bylipopolysaccharide: major histocompatibility complexclass II molecule dependency and cytokine release.Infect Immun;61:5333–5338.

19. Frei R., Steinle J., Birchler T., Loeliger S., Roduit C.,Steinhoff D., Seibl R., B€uchner K., Seger R., Reith W.,Lauener R.P. (2010) MHC class II molecules enhance Toll-like receptor mediated innate immune responses. PLoSONE;5:e8808.

20. Harton J.A., Van Hagen A.E., Bishop G.A. (1995) Thecytoplasmic and transmembrane domains of MHCclass II beta chains deliver distinct signals required forMHC class II-mediated B cell activation. Immu-nity;3:349–358.

21. Harton J.A., Bishop G.A. (1993) Length and sequencerequirements of the cytoplasmic domain of the A betamolecule for class II-mediated B cell signaling. J Immu-nol;151:5282–5289.

22. Kissner T.L., Ruthel G., Alam S., Mann E., Ajami D.,Rebek M., Larkin E, Fernandez S., Ulrich R.G, Ping S.,Waugh D., Rebek J. Jr, Kamal U., Saikh K.U. (2012)Therapeutic inhibition of pro-inflammatory signalingand toxicity to staphylococcal enterotoxin B by a syn-thetic dimeric BB-loop mimetic of MyD88. PLoSONE;7:e40773.

Supporting Information

Additional Supporting Information may be found in theonline version of this article:

Figure S1. Dose dependent inhibition of pro-inflammatorycytokine production by compound 4210 in primary PBMCsculture stimulated with SEB.

Figure S2. Pro-inflammatory cytokine inhibition by com-pound 1 and compound 4210 in primary culture ofPBMCs stimulated with SEB.

Figure S3. Dose dependent inhibition of pro-inflammatorycytokine production by compound 4210 in primary PBMCculture stimulated with LPS (F. tularensis or E. coli) or irra-diated B. mallei.

Figure S4. MyD88-specific signaling of SEAP reporterexpression inhibited by compound 4210.

Figure S5. Inhibition of SEAP reporter activity by com-pound 1 and compound 4210 stimulated with LPS (F. tu-larensis or E. coli) or B. mallei.



structure‐based design and synthesis of a small molecule

Documents