Fig 1.
The Ugi coupling reaction involves the combination of isocyanide (a), aldehyde (b), amine (c) and carboxylic acid (d) components in a single reaction step to give an α-aminoacyl amide product, as exemplified above for the preparation of AZ126.
Fig 2.
Optimization stages of Ugi products as TLR4 agonists.
*Human TLR-4 NFκB SEAP reporter assays; protocol as described in Methods. Data are reported as pEC50, or negative log of the EC50 values. **Clint rh / hm represents intrinsic clearance in presence of rat hepatocytes (μL/min/106 cells) and human microsomes (μL/min/mg protein). As a general guide, a value of <10 indicates good metabolic stability while >30 suggests that the compound is likely to be rapidly metabolized in vivo. †LogD represents the lipophilicity, whereby D is the octanol/water distribution coefficient at pH7.4. Thus a drug molecule with LogD >4 might be considered to be very lipophilic and therefore more likely to suffer from rapid metabolic turnover, poor solubility and receptor promiscuity; whereas a LogD range of 1–3 is more likely to be associated with optimal physicochemical and drug-like properties. ††Solubility was measured by dissolution of solid samples in buffer at pH 7.4.
Fig 3.
Ugi reaction products with variation of side-chains.
*LogD represents the lipophilicity whereby D is the octanol/water distribution coefficient at pH7.4; whereby LogD 1–3 is considered optimal for drug-like properties. **Human TLR-4 NFκB SEAP reporter assay; protocol as described in Methods (N.B. pEC50 = –Log10(EC50), thus values of 6 and 9 represent EC50 values of 1 μM and 1 nM (0.001 μM) respectively. ***NA = not available.
Fig 4.
Ugi compounds preferentially activate TLR4-HEK transfectants expressing human MD2.
A-B) HEK293 cells transfected with human TLR4/MD2 or mouse TLR4/MD2 (InVivogen) were stimulated with TLR4 agonists including MPL, PHAD, and Ugi compounds at 1 μg/ml to 4 ng/ml and with LPS at 25 ng/ml to 64 pg/ml for 24 h. SNs were analyzed for SEAP content by the QUANTI-Blue assay (InVivogen) and analysis at 650 nm by spectrophotometer. AUC (area under the curve) values were calculated as in Methods. No AUC value was calculated for LPS since it was utilized with a different dose range. C) HEK transfectants with cis- or trans-species expression of TLR4, MD2, and CD14 were stimulated with Ugi compounds in a dose response range of 9 μM to 1 nM or with LPS (boxes) at 75 ng/ml to 5 pg/ml for 48 h. Data was obtained as Vmax calculated over the initial linear portion of the kinetic absorbance measurements. The AUC (area under the curve) value for the dose response curve of Vmax vs log10 (molar units) was calculated for each compound using Trapezoidal rule.
Fig 5.
Ugi compounds activate guinea pig TLR4/MD2-HEK transfectants.
HEK293 cells underwent transient co-transfection with expression vectors for guinea pig TLR4, MD2, and CD14 along with SEAP reporter plasmid. Transfectants were stimulated with TLR4 agonist compounds at 10−5 to 10−10 M for 20 h. The plates were developed with the addition of DDAO phosphate as substrate and then read on a plate reader at 620 nm. Data were analyzed as % of the maximum 1 μM Ultrapure (0111:B4) LPS activation signal.
Fig 6.
The Ugi compound-induced cytokine profile is similar to MPL and LPS in human PBMCs but not mouse splenocytes.
PBMCs were isolated from three healthy volunteers, while murine splenocytes were isolated from three naïve BALB/c mice. TLR4 agonists were used at 1 and 0.2 μg/ml doses to stimulate human PBMCs and murine splenocytes for 24 hrs. SNs were analyzed via human and mouse MILLIPLEX kits. Data are expressed as the means of 3 donors/mice at the 1 μg/ml dose of TLR4 agonist and are representative of 2 studies. Raw data are presented in S3 Fig.
Fig 7.
Activation of cytokine gene expression by Ugi compounds in GP PBMCs.
PBMCs were isolated from three naïve GPs. TLR4 agonists were used at 1 and 0.2 μg/ml doses to stimulate GP PBMCs for 24 hrs. SNs were analyzed via GP Affymetrix QuantiGene Plex assay. Data are expressed as the means of 3 GPs at the 1 μg/ml dose of TLR4 agonist and are representative of 2 studies. Raw data are presented in S4 Fig.
Fig 8.
Secreted cytokine profile of rat PBMCs stimulated by AZ618.
PBMCs were isolated from three naïve Sprague-Dawley rats. TLR4 agonists were used at 1 and 0.2 μg/ml doses to stimulate human PBMCs and murine splenocytes for 24 hrs. SNs were analyzed via rat MILLIPLEX kits. Data are expressed as the means of 3 rats at the 1 μg/ml dose of TLR4 agonist and are representative of 2 studies. Raw data are presented in S5 Fig.
Fig 9.
Species-specific activation of IL-6, TNF-α, and chemokine expression in primary cells.
PBMCs were isolated from peripheral blood from human healthy volunteers, cynomolgus macaques, Hartley guinea pigs, ferrets, New Zealand white rabbits, Sprague-Dawley rats, and cotton rats, and splenocytes were isolated from BALB/c mice, n = 3 for each species, except n = 2 for cotton rats. TLR4 agonists were used at 1 μg/ml to stimulate primary cells of all species for 24 hrs. For human, cyno, mouse, and rat, SNs were analyzed via human, NHP, mouse, and rat MILLIPLEX kits. For guinea pig, ferret, rabbit, and cotton rat, cell lysates were analyzed via Affymetrix QuantiGene Plex kits. No one chemokine analyte was available in all 8 species-specific assays, so data were collected for any one of three different chemokines, MCP-1, MIP-1β, and RANTES. Data are expressed as mean compound % difference from the mean MPL value for that species and cytokine with associated 90% confidence interval (CI) listed below. A compound is considered significantly higher than MPL if the lower bound of the CI is higher than 0% (p value < 0.05); such values are in bold. Increasing shades of green and purple indicate increasingly higher and lower respective values than values achieved by MPL stimulation. Raw data are presented in S6 Fig.
Fig 10.
Ugi compounds activate mouse BMDCs that express human TLR4/MD2.
Bone marrow cells were isolated from femur/tibia bones of w.t. C57BL/6, TLR4/MD2 KO, and hTLR4/MD2 KIKO mice and cultured with GM-CSF. BMDCs were harvested at day 9, washed, counted, and stimulated in vitro for 24 h with 1 μg/ml TLR4 agonists and 2 μM R848. SNs were analyzed via mouse MILLIPLEX kits. Data are expressed as the means of 2 mice and are representative of 2 studies.
Fig 11.
AZ618 demonstrates in vivo activity in hTLR4/MD2 KIKO mice.
C57BL/6 mice (6–12 per group) were immunized IP with 50 μg MPL or AZ618 and then were bled at 2 h via retro-orbital route and at 6 h via cardiac puncture. Serum was analyzed by multiplex Millipore kits for cytokine content. Presented are individual results with a bar representing the group geometric mean with SD. Open circles represent wild-type C57BL/6 mice; X’s represent animals deficient for TLR4 and MD2; red circles represent TLR4/MD2 -/- mice transgenic for human TLR4/MD2. Statistical comparisons are between C57BL/6 w.t. and hTLR4/MD2 KIKO groups. Statistical test was one-way ANOVA performed with family-wise error rate to adjust multiple comparisons; ns, p > 0.05; *, p < 0.05; **, p < 0.01; ***, p < 0.001.
Fig 12.
Change of conformation in a flexible loop region between residues 120 and 129 of human MD-2 upon binding to LPS ligand and dimerization with a second TLR4/MD-2 unit.
Bound and unbound PDB structures were superimposed. The loop region is shown as stick representation (cyan–unbound; yellow–bound), with key residues colored separately (blue–unbound Phe126; magenta–bound Phe126; red–bound Tyr131; green–bound 130 and 132 residues). Bound Phe126 and Tyr131 are additionally rendered with semi-transparent spheres. The rest of MD-2 is shown as backbone trace and only for the unbound structure (cyan). Solid spheres show LPS ligand (orange), TLR4 of the first unit (chain A, green) and TLR4 of the second unit (grey). A) zoom out view with the entire LPS ligand; B) zoom in view with the second TLR4 unit hidden in order to provide better visibility for the conformation of the flexible loop of MD-2.
Fig 13.
Representative docking models.
Ugi ligand is green, other key residues—as labeled on the Fig The rest of MD-2 is yellow backbone trace with the flexible loop region shown as sticks. A) Docking mode where Ugi compound is predicted to interact with MD-2 residues Tyr131 and Phe126 (best-scored model for AZ617 docked into human receptor). B) Alternative docking mode where Ugi compound is deep inside the binding pocket away from Tyr131 and dimerization interface (best-scored model for AZ617 and the mouse receptor). C) Comparison of the docked Ugi compound and crystallographically solved Neoseptin-3 ligand in complex with the mouse receptor (model 7 from AZ161-mouse docking). MD-2 coordinates from both complexes were superimposed. AZ161—green sticks, Neoseptin-3—orange sticks.