David J. St. Jean

Mitigating Heterocycle Metabolism in Drug Discovery

■ INTRODUCTION In the past few decades, drug metabolism research has played an ever increasing role in the design of drugs. In vitro metabolism assays have become an integral part of the routine profiling of compounds made in drug discovery. The data from these assays have allowed medicinal chemists to focus their efforts on compounds with improved metabolic stability. Detailed metabolite identification studies are also done more routinely, which provide information on how to strategically replace or block metabolically labile sites. Additionally, in vivo PK studies are regularly conducted in drug discovery, which helps to build in vitro−in vivo PK relationships. The positive influence that these advances in PKDM sciences have had on drug discovery is reflected in the fact that fewer drug candidates fail in the clinic for PKDM related issues. This suggests that medicinal chemists are successfully integrating the data generated by their PKDM colleagues into the design of compounds with fewer metabolic liabilities. Extensive data from metabolism studies have allowed medicinal chemists to develop general principles for reducing compound metabolism. These methods include, but are not limited to, reducing lipophilicity, altering sterics and electronics, introducing a conformational constraint, and altering the stereochemistry of their compounds. While no single method is able to solve every metabolic problem, these principles do give medicinal chemists guidance on how to improve the metabolic liabilities of their compounds. If the specific site of metabolism is known, medicinal chemists block the site, typically with a fluorine, or replace the metabolically labile group with a bioisostere. While several authors have reviewed these techniques for reducing metabolism, there is no review that summarizes different approaches to improving the metabolic stability of heterocycles. In this review, we summarize examples where changes were made at or near the heterocycle to improve metabolic stability. By summarizing these examples, we hope to provide a useful guide to medicinal chemists as they attempt to improve the metabolic profile of their own heterocyclic compounds. The majority of the examples that are included in this review came from searching the online open access database CHEMBL. In addition to having pharmacology data on compounds from the medicinal chemistry literature, CHEMBL has over 120 000 points of data on the ADMET properties of compounds. With the help of the visualization software Spotfire, we were able to cull examples from the CHEMBL ADMET data that focused on heterocycles. We also identified examples from papers that cite leading reviews in the drug metabolism field and were present in other recent reviews on drug metabolism. The main criteria that we placed on the examples selected for this review was that the change made to improve metabolism had to occur at or near the heterocycle and nowhere else on the molecule. This allowed us to eliminate examples where a change made to a compound away from the heterocycle may have influenced the metabolism. The data that we included in this review is predominantly from in vitro microsomal stability studies. However, we have included some data from bioactivation studies and in vivo PK studies to provide additional information about the overall metabolic profile. In several instances, the compound with the improved metabolic profile also became the lead compound in the paper, so we felt that including the data on the intended target was informative even though this is not a discussion point for the review. Of course, in some examples when the heterocycle was modified to improve metabolic stability, the activity at the intended biological target diminished. However, we felt that these examples of improved metabolic stability would still be of value to the reader. In the discussion below, we have organized the review by first discussing saturated heterocycles and then heteroaromatic compounds. Within each section we have organized the discussion by ring size.

Antidiabetic effects of glucokinase regulatory protein small-molecule disruptors

Glucose homeostasis is a vital and complex process, and its disruption can cause hyperglycaemia and type II diabetes mellitus. Glucokinase (GK), a key enzyme that regulates glucose homeostasis, converts glucose to glucose-6-phosphate in pancreatic β-cells, liver hepatocytes, specific hypothalamic neurons, and gut enterocytes. In hepatocytes, GK regulates glucose uptake and glycogen synthesis, suppresses glucose production, and is subject to the endogenous inhibitor GK regulatory protein (GKRP). During fasting, GKRP binds, inactivates and sequesters GK in the nucleus, which removes GK from the gluconeogenic process and prevents a futile cycle of glucose phosphorylation. Compounds that directly hyperactivate GK (GK activators) lower blood glucose levels and are being evaluated clinically as potential therapeutics for the treatment of type II diabetes mellitus. However, initial reports indicate that an increased risk of hypoglycaemia is associated with some GK activators. To mitigate the risk of hypoglycaemia, we sought to increase GK activity by blocking GKRP. Here we describe the identification of two potent small-molecule GK–GKRP disruptors (AMG-1694 and AMG-3969) that normalized blood glucose levels in several rodent models of diabetes. These compounds potently reversed the inhibitory effect of GKRP on GK activity and promoted GK translocation both in vitro (isolated hepatocytes) and in vivo (liver). A co-crystal structure of full-length human GKRP in complex with AMG-1694 revealed a previously unknown binding pocket in GKRP distinct from that of the phosphofructose-binding site. Furthermore, with AMG-1694 and AMG-3969 (but not GK activators), blood glucose lowering was restricted to diabetic and not normoglycaemic animals. These findings exploit a new cellular mechanism for lowering blood glucose levels with reduced potential for hypoglycaemic risk in patients with type II diabetes mellitus.

Discovery of a Potent, Orally Active 11β-Hydroxysteroid Dehydrogenase Type 1 Inhibitor for Clinical Study: Identification of (S)-2-((1S,2S,4R)-Bicyclo[2.2.1]heptan-2-ylamino)-5-isopropyl-5-methylthiazol-4(5H)-one (AMG 221)

Thiazolones with an exo-norbornylamine at the 2-position and an isopropyl group on the 5-position are potent 11beta-HSD1 inhibitors. However, the C-5 center was prone to epimerization in vitro and in vivo, forming a less potent diastereomer. A methyl group was added to the C-5 position to eliminate epimerization, leading to the discovery of (S)-2-((1S,2S,4R)-bicyclo[2.2.1]heptan-2-ylamino)-5-isopropyl-5-methylthiazol-4(5H)-one (AMG 221). This compound decreased fed blood glucose and insulin levels and reduced body weight in diet-induced obesity mice.

Structural characterization and pharmacodynamic effects of an orally active 11beta-hydroxysteroid dehydrogenase type 1 inhibitor.

11β‐Hydroxysteroid dehydrogenase type 1 regulates glucocorticoid action and inhibition of this enzyme is a viable therapeutic strategy for the treatment of type 2 diabetes and the metabolic syndrome. Here, we report a potent and selective 11β‐hydroxysteroid dehydrogenase type 1 inhibitor with a binding mode elucidated from the co‐crystal structure with the human 11β‐hydroxysteroid dehydrogenase type 1. The inhibitor is bound to the steroid‐binding pocket making contacts with the catalytic center and the solvent channel. The inhibitor binding is facilitated by two direct hydrogen bond interactions involving Tyrosine183 of the catalytic motif Tyr‐X‐X‐X‐Lys and Alanine172. In addition, the inhibitor makes many hydrophobic interactions with both the enzyme and the co‐factor nicotinamide adenine dinucleotide phosphate (reduced). In lean C57BL/6 mice, the compound inhibited both the in vivo and ex vivo 11β‐hydroxysteroid dehydrogenase type 1 activities in a dose‐dependent manner. The inhibitory effects correlate with the plasma compound concentrations, suggesting that there is a clear pharmacokinetic and pharmacodynamic relationship. Moreover, at the same doses used in the pharmacokinetic/pharmacodynamic studies, the inhibitor did not cause the activation of the hypothalamic–pituitary–adrenal axis in an acute mouse model, suggesting that this compound exhibits biological effects with minimal risk of activating the hypothalamic–pituitary–adrenal axis.

Discovery of a Calcimimetic with Differential Effects on Parathyroid Hormone and Calcitonin Secretion

Calcimimetics are positive allosteric modulators to the calcium-sensing receptor (CaSR). Activation of the CaSR inhibits the secretion of parathyroid hormone (PTH), stimulates the secretion of calcitonin, and decreases serum calcium (Ca2+). Cinacalcet, a second-generation calcimimetic, is used therapeutically to control PTH in patients with chronic kidney disease who are on dialysis with secondary hyperparathyroidism. A calcimimetic that displays increased separation of PTH versus Ca2+ lowering in patients would potentially allow the use of calcimimetics to treat patients in earlier stages of renal disease because hypocalcemia can develop in this population. Toward this end, we developed a third-generation calcimimetic, determined the molecular pharmacological properties of it using an operation model of allosteric modulation/agonism, and measured the compound effects on PTH, serum ionized Ca2+, and calcitonin levels in 5/6 nephrectomized rats. We found the new molecule effectively reduced PTH levels without promoting calcitonin secretion or hypocalcemia. Furthermore, our third-generation molecule was less efficacious at promoting calcitonin secretion from human thyroid carcinoma cells compared with 3-(2-chlorophenyl)-N-((1R)-1-(3-methoxyphenyl)ethyl)-1-propanamine (R-568), a first-generation calcimimetic. These data provide evidence that calcimimetics with increased potency can be used to lower PTH without production of significant hypocalcemia because the threshold for inhibition of PTH secretion is much lower than the threshold for calcitonin secretion.

Small Molecule Disruptors of the Glucokinase–Glucokinase Regulatory Protein Interaction: 3. Structure–Activity Relationships within the Aryl Carbinol Region of the N-Arylsulfonamido-N′-arylpiperazine Series

We have recently reported a novel approach to increase cytosolic glucokinase (GK) levels through the binding of a small molecule to its endogenous inhibitor, glucokinase regulatory protein (GKRP). These initial investigations culminated in the identification of 2-(4-((2S)-4-((6-amino-3-pyridinyl)sulfonyl)-2-(1-propyn-1-yl)-1-piperazinyl)phenyl)-1,1,1,3,3,3-hexafluoro-2-propanol (1, AMG-3969), a compound that effectively enhanced GK translocation and reduced blood glucose levels in diabetic animals. Herein we report the results of our expanded SAR investigations that focused on modifications to the aryl carbinol group of this series. Guided by the X-ray cocrystal structure of compound 1 bound to hGKRP, we identified several potent GK-GKRP disruptors bearing a diverse set of functionalities in the aryl carbinol region. Among them, sulfoximine and pyridinyl derivatives 24 and 29 possessed excellent potency as well as favorable PK properties. When dosed orally in db/db mice, both compounds significantly lowered fed blood glucose levels (up to 58%).

Further Studies with the 2-Amino-1,3-thiazol-4(5H)-one Class of 11β-Hydroxysteroid Dehydrogenase Type 1 Inhibitors: Reducing Pregnane X Receptor Activity and Exploring Activity in a Monkey Pharmacodynamic Model

A series of compounds containing the 2-amino-1,3-thiazol-4(5H)-one core were found to be potent inhibitors of the enzyme 11beta-hydroxysteroid dehydrogenase type 1 (11beta-HSD1). One of our lead compounds from this series activated the human nuclear xenobiotic receptor, pregnane X receptor (PXR). To try and mitigate the PXR activity, we prepared analogues of our lead series that contained polar groups on the right-hand side of the thiazolone. Several analogues containing amides or alcohols appended to the C-5 position of the thiazolone showed a significant reduction in PXR activity. Through these structure-activity efforts, a compound containing a tert-alcohol group off the C-5 position, analogue (S)-33a, was found to have an 11beta-HSD1 Ki = 35 nM and negligible PXR activity. Compound (S)-33a was advanced into a pharmacodynamic model in cynomolgus monkeys, where it inhibited adipose 11beta-HSD1 activity after being orally administered.

Small Molecule Disruptors of the Glucokinase–Glucokinase Regulatory Protein Interaction: 1. Discovery of a Novel Tool Compound for in Vivo Proof-of-Concept

Small molecule activators of glucokinase have shown robust efficacy in both preclinical models and humans. However, overactivation of glucokinase (GK) can cause excessive glucose turnover, leading to hypoglycemia. To circumvent this adverse side effect, we chose to modulate GK activity by targeting the endogenous inhibitor of GK, glucokinase regulatory protein (GKRP). Disrupting the GK-GKRP complex results in an increase in the amount of unbound cytosolic GK without altering the inherent kinetics of the enzyme. Herein we report the identification of compounds that efficiently disrupt the GK-GKRP interaction via a previously unknown binding pocket. Using a structure-based approach, the potency of the initial hit was improved to provide 25 (AMG-1694). When dosed in ZDF rats, 25 showed both a robust pharmacodynamic effect as well as a statistically significant reduction in glucose. Additionally, hypoglycemia was not observed in either the hyperglycemic or normal rats.

Small molecule disruptors of the glucokinase-glucokinase regulatory protein interaction: 2. Leveraging structure-based drug design to identify analogues with improved pharmacokinetic profiles.

In the previous report , we described the discovery and optimization of novel small molecule disruptors of the GK-GKRP interaction culminating in the identification of 1 (AMG-1694). Although this analogue possessed excellent in vitro potency and was a useful tool compound in initial proof-of-concept experiments, high metabolic turnover limited its advancement. Guided by a combination of metabolite identification and structure-based design, we have successfully discovered a potent and metabolically stable GK-GKRP disruptor (27, AMG-3969). When administered to db/db mice, this compound demonstrated a robust pharmacodynamic response (GK translocation) as well as statistically significant dose-dependent reductions in fed blood glucose levels.

Synthesis and Evaluation of the Metabolites of AMG 221, a Clinical Candidate for the Treatment of Type 2 Diabetes.

All eight of the major active metabolites of (S)-2-((1S,2S,4R)-bicyclo[2.2.1]heptan-2-ylamino)-5-isopropyl-5-methylthiazol-4(5H)-one (AMG 221, compound 1), an inhibitor of 11β-hydroxysteroid dehydrogenase type 1 that has entered the clinic for the treatment of type 2 diabetes, were synthetically prepared and confirmed by comparison with samples generated in liver microsomes. After further profiling, we determined that metabolite 2 was equipotent to 1 on human 11β-HSD1 and had lower in vivo clearance and higher bioavailability in rat and mouse. Compound 2 was advanced into a pharmacodynamic model in mouse where it inhibited adipose 11β-HSD1 activity.