Figure 20.25 • Primary amino acid sequences of human amylin, rat amylin, and pramlintide. Charge symbols above the set of sequences indicate charges of the predominant molecular form of the side chain (the free N-terminal amino group is also indicated) at physiological pH (7.4): positive (+), neutral (0), or near-equal (=). Underlines indicate sites of amylin hydrolysis by insulin-degrading enzyme (IDE).

the periphery, if any, remain poorly characterized.81 Unsurprisingly, though, the appetite-suppressing actions, and benefits of pramlintide and amylin agonists in general, are receiving significant clinical and scientific attention.

Pramlintide has a relatively short pharmacokinetic halflife of about 50 minutes. One conversion product, des-Lys1-pramlintide, retains full agonist activity and exhibits an elimination time-course similar to that of pramlintide itself. Amylin is degraded by IDE at three sites (as indicated in Fig. 20.25); in vivo administration of IDE inhibitors increases toxicities of the same nature as those associated with excessive endogenous or exogenously administered amylin.82 Pramlintide possesses amino acid residue pairs identical with those of amylin at each of the IDE cleavage sites, thus this enzyme probably governs the elimination rates of pramlintide and des-Lys1-pramlintide.

Despite the fact that pramlintide is a relatively large and clearly nonnative polypeptide, systemic immunosensitivity has not proven to be an issue of clinical concern, even upon long-term pramlintide pharmacotherapy.

Incretin System Modulating Agents: Incretin Mimetics and Dipeptidyl Peptidase Type 4 Inhibitors

Several polypeptide hormones have been identified that are synthesized in the cells of the upper intestines and act to modulate carbohydrate metabolism. Among these are the in-cretins, a term first coined by La Barre in 1932.83 The "in-cretin effect" refers to insulin secretion stimulated by orally administered glucose independent of any increase in blood glucose. Eventually, two incretins were identified, in 1971 and 1985, respectively: glucose-dependent insulinotropic peptide (GIP) and GLP-1.84 GIP was originally named gastric inhibitory peptide by its discoverers,85 but it was later renamed in such a way as to retain the acronym while better reflecting its physiological roles.

The gene encoding proglucagon (PG), besides being expressed in pancreatic islet a cells (and in some neurons in the hypothalamus and brainstem, and certain taste cells of the tongue), is also expressed in endocrine cells of the lower intestinal mucosa. From PG, gut L cells produce the peptides glicentin, oxyntomodulin, intervening peptide-2, and GLP-1. PG(72-108) is equivalent to GLP-1(1-37), and is derived from PG via the action of the convertase PC1/3. GLP-1 (1-37) is further shortened, apparently also by PC1/3, to GLP-1(7-37) or GLP-1(7-36)amide (Fig. 20.26); at least 80% of "GLP-1" released into the bloodstream is the latter. Secretion occurs upon ingestion of lipid- or carbohydrate-containing foodstuffs.86-88 Following their release, both bioactive forms of GLP-1 are very short-lived (plasma t/2 —1-2 min): DPP-4 or DPP-IV rapidly cleaves two residues from the amino terminus, generating GLP-1(9-37) and GLP-1(9-36)amide, respectively.88 Release of

Tyr-Ala-Glu-Gly-Thr-Phe-lle-Ser-Asp-Tyr-Ser-lle-Ala-Met-Asp-Lys-lle-His-Gln-GIn-Asp-Phe-Val-Asn-Trp-Leu-Leu-Ala-Gln-Lys-Gly-Lys-Lys-Asn Asp-Trp-Lys-His-Asn-lle-Thr-GIn OH


H -His-Ala-Glu-Gly-Thr-Phe-Thr-Ser-Asp-Val-Ser-Ser-Tyr-Leu-Glu-Gly-Gln-Ala-Ala—N'^h—Glu-Phe-lle—Ala-Trp-Leu-Val-Lys-Gly-Arg-Gly- OH 7 8 | 15 A 18 A a20 H O 27* A 30 * 36 37

DPP H2NyNH2 etc Gly-Arg NH2 GLP-1 (7-36)-amide

itide i

H -His-Gly-Glu-Gly-Thr-Phe-Thr-Ser-Asp-Leu-Ser-Lys-Gln-Met-Glu-Glu-Glu-Ala-Val-N-^—Leu-Phe-lle-Glu-Trp-Leu-Lvs-Asn-Gly-Gly-Pro-Ser-Ser-Gly-Ala-Pro-Pro-Prp-Ser- NH2 7 8 15 18 20 ° 27 30 A 36


H -His-Ala-Glu-Gly-Thr-Phe-Thr-Ser-Asp-Val-Ser-Ser-Tyr-Leu-Glu-Gly-Gln-Ala-Ala——Glu-Phe-lle—Ala-Trp-Leu-Val-Arg-Gly-Arg-Gly- OH


taspoglutide j jf H,c CH„

H -His—Alb-Glu-Gly-Thr-Phe-Thr-Ser-Asp-Val-Ser-Ser-Tyr-Leu-Glu-Gly-Gln-Ala-Ala-N'^r—Glu-Phe-lle—Ala-Trp-Leu-Val-Lys-Aib-Arg NH2 Ajb _


H -His-Gly-Glu-Gly-Thr-Phe-Thr-Ser-Asp Val-Ser-Ser-Tyr-Leu-Glu-Gly-Gln-Ala-Ala-N'^v—Glu-Phe-lle-Ala-Trp-Leu-Val-Lys-Gly-Arg -HSA


Figure 20.26 • Primary sequences of endogenous human incretins and incretin mimetic pharmacotherapeutic agents. Arrows indicate a peptide bond cleaved by DPP-4 catalysis, carets (A) indicate sites of cleavage by neutral endopeptidase 24.11 (neprilysin) catalysis (double-caret-marked bonds are likely more prominently cleaved). Underscored residues in the structure of exenatide indicate nonconservative substitutions versus GLP-1 or GIP. The numbering for GLP-1 is "slipped" by six residues; the corresponding numberings are given below the GIP structure for reference. (HSA, human serum albumin.)

the bioactive forms of GLP occurs within minutes of food intake, even though L cells are found predominantly in the ileum and caecum, and considerable evidence indicates that some combination of neural or endocrine signaling mediates this response. GLP-1 acts via specific GPCRs in the portal vein to trigger vagal afferents that, mediated by neuronal pathways within the brain, generate efferent signals stimulating pancreatic insulin release and inhibiting glucagon release.89 GPCRs for GLP-1 (GLP-1Rs) are also expressed in pancreatic islet 3 cells, as well as in the GI tract, kidney, vagus nerve, and neurons in areas of the brain, including the hypothalamus and hindbrain; in ¡-islet cells, agonist binding to GLP-1Rs increases the biosynthesis of insulin, as well as of glucokinase and the GLUT2 glucose transporter.90

Structure-activity studies indicate that the C-terminal portions of the GLP-1 peptides provide an important component of receptor binding, whereas the N-terminal segment is necessary for receptor activation.91 Accordingly, once DPP-4 removes the N-terminal dipeptide from either GLP-1(7-37) or GLP-1(7-36)NH2, agonist activity is lost. GLP-1(9-36)NH2 finds some use as a pharmacological tool for its antagonist activity at GLP-1Rs.

GIP is synthesized via the action of a prohormone con-vertase on a 153-amino-acid proprotein coded by a specific gene in gut K cells, which are found mainly in the mucosa of the duodenum and jejunum.84 In contrast to the mostly indirect stimulation of GLP-1 release, GIP secretion is mainly triggered directly by the enteral presence of glucose and lipids, providing what amounts to an early warning system for the pancreas. The bioactive GIP(1-42) is rapidly converted (plasma ^ —5-7 min) to inactive GIP(3-42). Like the two GLP-1 hormones, this inactivation is also effected via DPP-4 catalyzed cleavage of a dipeptide from the amino ter-minus.92 GIP activates insulin secretion via binding to seven-transmembrane-segment GPCRs on islet 3 cells, distinct receptors from those activated by GLP-1(7-37) and GLP-1(7-36)NH2.88 GIP-stimulated insulin secretion only occurs when blood glucose is substantively elevated, and receptor knockout studies in mice indicate that GIP receptors (GIPRs, which are also present in brain, pituitary, adipocytes, upper GI tract, adrenal cortex, and bone) are required for a proper gluconormative response to hyper-glycemia. Notably, in type 2 diabetics, exogenously administered (iv) GIP cannot facilitate glucose-stimulated insulin secretion as it does in healthy patients.

Relatively recently, compelling direct evidence from rodent and in vitro experimentation and some indirect but inconclusive evidence from human clinical studies of in-cretin mimetics (see discussion that follows) and DPP-4 inhibitors suggests that incretins may exert preservative, proliferative, and even neogenerative effects on ¡-cell mass and ¡-cell populations, thereby arresting—and potentially even reversing—the progressive declines that have, heretofore, been inevitable in type 2 diabetes.90,93 If so, a truly new age in the treatment of this common and insidious disorder has begun, particularly when technology allows reliable detection of prediabetic states amenable to prophylactic treatment.

Exenatide, at the time of this writing, is currently the only incretin mimetic approved (2005) for the U.S. market (Byetta, Amylin Pharmaceuticals). Exenatide must be administered parenterally (subcutaneous [sc]); injection is typically made in the abdomen about an hour before the first and last meals of the day. Exenatide is synthetically produced exendin-4, a 39-residue hormone (see Fig. 20.26) present in the venom of the Gila monster. Exenatide's actions mimic those of GLP-1; exenatide differs from GLP-1 in 14 of its first 30 amino acids (the length of GLP-1(7-36)NH2), although at least four of these substitutions may be regarded as conservative. Exenatide is not, however, a DPP-4 substrate, by virtue of the presence of a glycine rather than an alanine in the second position, and partly for this reason has a greatly increased serum half-life of 2 to 3 hours94; some of the other amino acid changes almost certainly also render exenatide less susceptible to neutral endopeptidase 24.11 (NEP-24.11),95 which is a membrane-bound zinc metallopeptidase also known as neprilysin, and which may account for as much as half of the elimination of GLP-1(7-37) and GLP-1(7-36)NH2. Although approved as adjunctive therapy (with metformin, or with a sulfonylurea, or with metformin + a sulfonylurea, or with metformin + a thiazolidinedione), in fact none of these other medicinals are necessary for exenatide to produce its clinically beneficial effects. Exenatide administration alone in pharmacological doses restores the early stage insulin release, loss of which is a hallmark of type 2 diabetes, and increases the robustness of the second-stage insulin release. Premeal exenatide injections (5 or 10 fMg) also suppress glucagon release following a meal (but do not affect normal hypoglycemia-stimulated glucagon release), slow gastric emptying, induce satiety, and promote normalcy of postprandial glucose uptake while suppressing hepatic glucose release. Moreover, the glucose dependency of exenatide-stimulated insulin secretion means that this drug cannot cause abnormal hypoglycemic episodes, although this can, of course, happen when the drug is adjunctive to a sulfonylurea. Although there are GLP-1 receptors expressed in hindbrain and hypothalamus that are probably involved in mediating exenatide-induced satiety effects, evidence suggests that direct agonist action on these receptors by circulating exenatide is likely not involved.96

Liraglutide is GLP-1(1-37) modified at Lys26: the e-amino group is acylated with an (N-hexadecanoyl)glutam-y-yl moiety (Fig. 20.26) that noncovalently but avidly binds to human serum albumin. (Also, Lys34 is conservatively replaced with Arg.) These changes provide for a greatly extended serum of —10 to 15 hours after sc administration (0.6-1.8 mg), allowing for once-daily dosing.97-99 Liraglutide (Victoza, Novo Nordisk) is awaiting approval by the FDA, EMEA, and the Koro-sho (Japan).

Albiglutide (also known as naliglutide, GSK 716155) is a sequential dimer of GLP-1(7-36)(Alas^Gly) covalently linked to modified human serum albumin (i.e., it is a "fusion protein" with modified HSA). This molecule is currently well into phase III clinical trials. Although albiglutide, if marketed (Syncria, GlaxoSmithKline), will—like exenatide—be administered by injection, the half-life and duration of 6 to 7 days100 would allow for weekly administration.

Taspoglutide (R1583, Roche) is GLP-1 modified at residues 8 and 35 (see Fig. 20.26) with a-methylalanine in place of the Ala and Gly, respectively. These modifications greatly retard degradation by DPP-IV and neutral endopep-tidase 24.11, respectively (see preceding discussion for exe-natide). The phase III trials, which commenced in 2008, are being conducted with a zinc-based preparation, prolonging the duration in a manner similar to zinc-containing insulin preparations, so that once-weekly injections of 20 to 30 mg will be possible.

Besides the aforementioned list of products, a number of other incretin mimetics are being created and developed. A long-acting version of exenatide ("exenatide LAR") consisting of microspheres of a poly(lactide-coglycolide) polymeric matrix has been devised, from which exenatide is slowly released from the injection site through diffusion and erosion.101 An adenoviral vector has also been created for exenatide "gene therapy," and was reported to successfully generate exenatide and modulate glucose control and other GLP-1-mediated responses in mice for at least 15 weeks following treatment102; other gene therapy approaches for delivering GLP-1 itself are also being explored. Boc5,103 a cyclobutanedicarboxylic acid (CAS Reg. No. 917569-14-3, structure still undisclosed), is reported to be a nonpeptide GLP-1 agonist exerting incretin effects in C57BL/6J and db/db mice. The variety of the developments described above provide some sense of the current level of interest in incretin mimetics among researchers and clinicians, interest undoubtedly based on perceived strength of promise. The relative merits of incretin receptor agonists as compared to the indirect incretin-enhancing approach achieved with inhibitors of DPP-IV (discussed next) or neutral endopepti-dase 24.11 largely remain to be discerned, which will likely happen during the coming decade or so.

DPP-4 or DPP-IV is a serine protease that selectively binds and hydrolyzes substrates that incorporate alanine (as in GLP(7-36)NH2, see earlier discussion and Fig. 20.26) or proline (as in neuropeptide Y and peptide YY) at the second position. Because there are many serine proteases, excellent selectivity is an essential design feature requisite to bestowing pharmacotherapeutic usefulness.104 In particular, minimizing inhibition of the closely related enzymes DPP-2, fibroblast activation protein (FAP), DPP-7 (also known as quiescent cell proline peptidase, QPP), DPP-8, and DPP-9 has proved to be crucial.105,106 These serine proteases are among a relatively few peptidases that can cleave at proline, and they do so on the carboxyl side. Human DPP-4 is a 766-amino-acid transmembrane glycoprotein and a member of peptidase clan SC, and more specifically a member of family S9, subfamily S9B. Clan SC serine proteases include a catalytic triad in the specific order Ser, Asp, His—different from the classical serine proteases such as trypsin (class SA: His, Asp, Ser) or subtilisin (class SB: Asp, His, Ser).

DPP-4 was first identified in 1967, and early inhibitors were created by relatively simple modifications of proline-based structures. After the later realization that this enzyme might likely be an important pharmacotherapeutic target, compound library screening provided inhibitors with a variety of new base structures, including the lead compound from which sitagliptin was created (Fig. 20.27). In due course, a significant number of x-ray crystal structures were generated, for complexes in which the catalytic pocket is occupied with various inhibitors, the first two reports occurring in

2003.107,108 This body of structural information very quickly grew to be sufficient to enable true de novo structure-based drug design; an excellent overview is provided by Kuhn et al.109 The catalytic site of DPP-4 is extracellular, and includes an a/S-hydrolase fold containing the Ser630-Asp708-His740 catalytic triad, and an eight-bladed S-propeller domain108 that participates in the binding of various known inhibitors, including sitagliptin.109 The S1 pocket within the protein, which accommodates proline (or, less favorably, alanine, as is found in the GLP-1 peptides and in GIP), is enclosed by the largely hydrophobic residues Tyr631, Val656, Trp662, Tyr666, and Val711, and among the reported x-ray structures of enzyme:inhibitor complexes the shape of this pocket changes very little, even for structurally very disparate inhibitors.

The discovery of sitagliptin at Merck preceded the availability of x-ray structures, and was accomplished by more classical (albeit highly refined) screening and lead-optimization approaches. Thornberry and Weber106 provide an informative account of this process. Although initially commencing with preclinical development of in-licensed compounds, extensive screening of a very large compound collection was also conducted c. 2001, from which effort these authors report that relatively few hits— and only three suitable leads—were gleaned. Sitagliptin was ultimately created (Fig. 20.27) from one of these leads, which was a S-amino acid derivative and thus unlike the a-amino-acid-based compounds divulged prior to this discovery. The in-licensed compounds were abandoned at about the same time because of toxicities that were found to be traceable to DPP-8 or DPP-9 inhibition; the existence of these enzymes, as well as of QPP (see previous discussion), were first reported circa this same time frame. Some characteristics of a few of the "way-station" compounds created during the lead optimization process that led to sitagliptin provide a sense of the process, particularly the large changes in target-level activity, biopharmaceutics characteristics, and deleterious effects profile, which can result from seemingly very subtle changes to the molecular structure. The necessity to monitor and optimize multiple characteristics simultaneously (such as target affinity and oral bioavailability) during the process of creating a marketable drug should also be readily apparent from this outline. Sufficient preclinical studies were completed for sitagliptin by January 2002 that it was moved to first priority for development. (Marketing approval was obtained from the FDA in late 2006, representing an impressively and exceptionally compressed timeline from screening hit identification to marketing.)

An x-ray structure of DPP-4 with bound sitagliptin later provided an understanding of the structure-affinity observa-tions.110 The trifluorophenyl moiety occupies the S1 pocket. The triazolopyrazine-ring-attached trifluoromethyl sub-stituent—where a buildup of electron density occurs—situ-ates near the positively charged side chain of Arg358; the protonated (cationic) form of the primary amino group forms a hydrogen bond with Tyr662, and salt bridges to the y-carboxylates of Glu205 and Glu206; the triazolopiperazine group engages the phenyl group of residue Phe357.

Other DPP-4 inhibitors are either on the market elsewhere (vildagliptin in the EU, Novartis [Galvus]), or are under regulatory review (alogliptin, Takeda; saxagliptin,

Saxagliptin StructureAlogliptin Bound Active Site Dpp

Figure 20.27 • Sitagliptin creation outline. The ^-amino acid-lead compound (1) was discovered by screening of a large library of compounds. Compound 2 bound with very high affinity to the enzyme but was not endowed with useful oral bioavailability. Changing the metabolically labile ethyl substituent in compound 3 to trifluoromethyl gave a compound (4) with acceptable bioavailability, albeit binding with considerably lower affinity than molecule 2. Relocating the fluorines on the phenyl ring improved the binding affinity but a cardiovascular side effect appeared in preclinical studies of 5 in dogs. Adding one additional fluorine increased the affinity and eliminated the worrisome preclinical side effect; molecule 6 was ultimately marketed as sitagliptin phosphate.

Figure 20.27 • Sitagliptin creation outline. The ^-amino acid-lead compound (1) was discovered by screening of a large library of compounds. Compound 2 bound with very high affinity to the enzyme but was not endowed with useful oral bioavailability. Changing the metabolically labile ethyl substituent in compound 3 to trifluoromethyl gave a compound (4) with acceptable bioavailability, albeit binding with considerably lower affinity than molecule 2. Relocating the fluorines on the phenyl ring improved the binding affinity but a cardiovascular side effect appeared in preclinical studies of 5 in dogs. Adding one additional fluorine increased the affinity and eliminated the worrisome preclinical side effect; molecule 6 was ultimately marketed as sitagliptin phosphate.

Bristol Myers Squibb [Onglyza]),111 or are well along in phase III studies (linagliptin, Boehringer Ingelheim). Three of these (Fig. 20.28) incorporate a nitrile (cyano) moiety: vildagliptin and saxagliptin a 2-cyanopyrrolidine, alogliptin a 2-cyanophenyl. In each case, the nitrile engages the enzyme at the same position as does the scissile bond in peptide substrates such as GLP-1 or GIP. The nitrile group is attacked by the hydroxyl of the catalytic serine (Ser630), forming a covalent imidate species (Fig. 20.29). Hydrolysis of this imidate occurs slowly; thus, these inhibitors may be described as slowly reversible or pseudoirreversible.

Sitagliptin phosphate is the 1:1 phosphoric acid salt of sitagliptin free base (i.e., (2R)-4-oxo-4-[3-(trifluoromethyl)-5,6-dihydro[1,2,4]triazolo[4,3-a]pyrazin-7(8H)-yl]-1-(2,4,5-trifluorophenyl)-2-butanamine), see Figure 20.30, and is marketed (Junuvia, 2006) in 25-, 50-, and 100-mg tablets. A combination product with metformin (Janumet, 2007) in two strengths (50 mg/500 mg and 50 mg/1,000 mg sitagliptin/metformin) is also available, and sitagliptin may also be prescribed with a thiazolidinedione, or possibly a sulfonylurea. The phosphate salt provides very high water solubility.112 The bioavailability of orally administered sitagliptin is —87%.113 The drug exhibits relatively low plasma protein

Was this article helpful?

0 0
Abolish Acne Today

Abolish Acne Today

This Book Below Will Show You Exactly What What You Need To Do To Finally Be A Success With Abolishing Acne!

Get My Free Ebook

Post a comment