Nimbus Discovery - 25 First Street, Suite 404 - Cambridge, MA 02141 USA - Phone: (617) 588-2655 - Fax: (617) 231-0652 - Email:[email protected] - Web:www.nimbusdiscovery.com Acetyl-CoA Carboxylase Inhibition by ND-630 Inhibits Fatty Acid Synthesis and Stimulates Fatty Acid Oxidation in Cultured Cells and in Experimental Animals Geraldine Harriman, Jeremy Greenwood*, Sathesh Bhat*, Liang Tong # , Ruiying Wang # , Debamita Paul # , Rosana Kapeller and H. James Harwood Jr. Nimbus Discovery, Inc., Cambridge, MA, USA; *Schrödinger, Inc., New York, NY, USA; # Columbia University, New York, NY, USA N O N O N O O O O CH 3 O OCH 3 OH OH OCH 3 0 12 5,000 Months Ideas Milestone Achieved Confirm binding mode (x-ray) 175 Total synthesized compound 8,000 nM <10 nM Potency ACC1, ACC2 1.3MM In vivo target engagement 16 10,000 225 <1 nM Rat DIO Studies in progress mg TGs per g Liver Tissue 0 2 4 3 10 30 Vehicle ND-630 (mg/kg) Lean DIO 0 100 150 200 250 300 0.39 1.56 6.25 25 100 Inhibitor (μM) mA 50 Probe + cpd + ACC2 Probe + cpd + buffer -787 -387 -187 13 213 1013 30 40 50 60 70 80 Temperature (°C) Fluorescence (dR) 413 613 -587 813 5 μM hBC 5 μM hBC + 50 μM Cp 5 μM hBC + 20 μM Soraphen 5 μM hBC + 100 μM Ligand 5 μM hBC + 50 μM Ligand 5 μM hBC + 20 μM Ligand 0.001 0.010 0.100 1.000 10.000 100.000 0 4 8 12 16 20 24 Time (hrs) Concentration (μM) IV PO_A 0.001 0.01 0.1 1 10 0 10000 20000 30000 40000 cpm ND-630 CP640186(3 μM) Inhibitor (μM) ED 50 = 0.14 mg/kg PO 0 10 20 30 40 50 60 70 Vehicle 0.3 1 mg/kg 3 10 30 DPMs/mg Tissue MED = 3 mg/kg PO 0.7 0.75 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 -60 0 60 120 180 240 RQ (±SEM) Vehicle 3mpk 10mpk 30mpk Time (min) Mitochondrion Acetyl -CoA ACC Inhibitor ACC1 Malonyl- CoA Malonyl- CoA Fatty acid synthesis Fatty acid oxidation ACC2 Hepatic Lipid Lowering Liver Targeted Liver Disease Inhibitor Mediated Effects Potential For Disease Modification Improvement in Insulin Sensitivity with Weight Neutrality Type 2 Diabetes Plasma Lipid Lowering Dyslipidemia ABSTRACT Inhibition of acetyl-CoA carboxylase (ACC) reduces fatty acid synthesis and stimulates fatty acid oxidation and has the potential to favorably affect the morbidity and mortality asso- ciated with obesity, diabetes, and fatty liver diseases. ACC exists as two tissue-specific isozymes; ACC1 present in li- pogenic tissues and ACC2 present in oxidative tissues. To achieve maximal effectiveness an ACC inhibitor should inhibit both isozymes. Most efforts to discover ACC inhibitors have focused on in- teractions within the carboxyltransferase (CT) domain of the enzyme active center. By contrast, our efforts have focused on the allosteric subunit dimerization site on the biotin car- boxylase (BC) domain where Soraphen interacts. This is also the same site where the phospho-containing motif of phos- pho-ACC binds to prevent dimerization and subsequent acti- vation of the enzyme. Using state-of-the-art structure-based drug design tech- niques together with crystal structures of the BC domain of human ACC2, we identified a unique series of small mole- cule allosteric inhibitors that bind to the Soraphen binding site and inhibit enzymatic activity. Development of this series has yielded deep structure-activity relationships, sub-nanomolar enzyme inhibition, functional activity in HepG2 and C2C12 cells and favorable drug-like properties leading to in vivo proof of concept. For example, the series representative ND-630 inhibits both human ACC1 and ACC2 (IC 50 = 2.0 nM), inhibits HepG2 cell fatty acid synthesis (EC 50 = 66 nM), stimulates C2C12 cell fat- ty acid oxidation (2-fold at 200 nM), inhibits rat hepatic fatty acid synthesis (ED 50 = 0.14 mg/kg) and stimulates rat whole body fatty acid oxidation (minimum effective dose 3 mg/kg). Together these observations suggest that allosteric ACC inhibition has the potential to favorably impact diabetes, obe- sity, fatty liver and lipid disorders. • A structure-based virtual screen and drug design approach utilizing WaterMap was used to identify allosteric inhibitors of ACC that bind to the BC domain. • Inhibitors were successfully optimized for excellent potency and drug-like properties in 12 months. • ND-630 demonstrated in vivo proof of concept in pharmaco- logically relevant models of target engagement (rat FASyn ED 50 = 0.14 mpk, PO & rat RQ MED = 3 mpk, PO). FIGURE 1: Acetyl CoA Carboxylase (ACC): Master Regulator of Fatty Acid Synthesis & Oxidation SUMMARY NEXT STEPS • Using this approach, a portfolio of nanomolar inhibitors with diverse functional-group driven bio-distribution patterns have been identified and are being utilized for a breadth of ACC- aligned indications. • Nimbus will report on the detailed pharmacology of ND-630 and other molecules in metabolic disease models, diabe- tes models and oncology models (cancer metabolism) in the future. • Beneficial effects on lipids, blood glucose, weight, potentially diabetes and CV risk. • Nimbus: FIRST small molecule allosteric inhibitor successfully targeting BC domain. FIGURE 2: Nimbus Allosteric Inhibitors Show Promise for Both Metabolic Disease & Fatty Liver Disease FIGURE 3: Previous Approaches Identified Inhibitors of the ACC CT Domain (ex. CP-640186) FIGURE 4: Nimbus’ Unique Approach Identified Potent ACC BC Domain Inhibitors Targeting Soraphen Binding Site FIGURE 5: Nimbus Approach has Delivered ACC Inhibitors with Demonstrated in vivo PoC in 12 Months • Most CT-domain inhibitor series identified exhibited poor drug-like properties. • Nimbus took an orthogonal approach and designed inhibitors that bound to the Soraphen binding site in the BC domain of the ACC complex. • A structure-based virtual screen of 1.3M lead-like molecules was performed utiliz- ing Schrodinger’s computational tools including WaterMap and Glide. This led to the identification of a high quality family of hit molecules with micromolar inhibitory effects against the enzyme. Co-crystal structures of members of this series in the human BC domain confirmed the model and our approach. • In an iterative design fashion over the next 12 month period, the potency of this fam- ily of hits were improved 1000x utilizing the computational model focusing on the re- placement or stabilization of high energy hydration sites within the Soraphen binding site. • Simultaneous to the potency improvements, drug-like properties were optimized to deliver Development Candidate quality molecules that are currently being profiled in chronic models of metabolic syndrome and diabetes. FIGURE 6: In vitro Assays Confirm Binding of Early Hits to BC Domain Prior to Obtaining Co-crystal Structures • Early in the program we utilized the Soraphen-TAMRA binding assay and BC domain thermal shift assays to provide us confidence that our inhibitors in fact were binding in the Soraphen binding site. • These tools were utilized until the first Nimbus inhibitor was co-crystalized in the hu- man BC domain. This co-crystal structure taught us our computational model was highly predictive and also provided additional information for WaterMap calculations. FIGURE 7: ND-630 Displays Excellent Drug-like Properties FIGURE 8: ND-630 in vivo Proof of Concept: Target Engagement in the Liver and Muscle TARGET POTENCY • ACC1 IC 50 = 2 nM • ACC2 IC 50 = 2 nM • HepG2 EC 50 = 66 nM • C2C12 FAOxn stim. 2x @ 200 nM ADMET • Low multispecies intrinsic clearance (human, mouse, rat, dog, monkey) • High solubility (>300 mM) • P450 inhib >50 mM, hERG >30 mM • Protein binding (98%) • DrugMatrixPanel 0/120 hits; 1000x window RAT PK (10 mg/kg, PO) • 37%F, Cmax 6 mM, Cl 33 mL/min/kg, AUC 3.4 mM.hr, T1/2 = 4.5h, Vss 2 L/kg • Liver exposure: 81 mM • Muscle exposure: 0.45 mM PHARMACOLOGY • Rat RQ MED = 3 mpk PO • Rat FASyn ED50 = 0.14 mpk, PO Soraphen-TAMRA Assay (Hit displaces soraphen) Thermal Shift Assay (Hit binds to the BC domain) Plasma PK Profile SD Rats IV (3 mpk, n=3, solution dose) PO_A (10 mpk, n=3, suspension dose) Hep-G2 cell FAsyn Inhibition [14C]acetate incorporation into FA) Target Engagement in Liver (Fatty acid synthesis inhibition) Target Engagement in Muscle (Respiratory quotient) Confirmatory Co-crystal Structures Soraphen A CP-640186