R E V I E W S
`
`A PROTON-PUMP INHIBITOR
`EXPEDITION: THE CASE HISTORIES OF
`OMEPRAZOLE AND ESOMEPRAZOLE
`
`Lars Olbe, Enar Carlsson and Per Lindberg
`
`Thirty years ago, disorders associated with inappropriate levels of gastric acid were a major
`problem for which treatment options were limited, and approaches to the control of gastric acid
`secretion were thus the focus of considerable drug discovery efforts. Here, we summarize how
`one such programme led to the development of the proton-pump inhibitor omeprazole (Losec,
`Prilosec), a conceptually new drug that proved clinically superior to previous antisecretory drugs
`in the treatment of acid-related disorders, and which became the world’s best-selling drug in the
`late 1990s. We then describe how the antisecretory and clinical effects were further improved by
`the development of esomeprazole (Nexium), a single enantiomer of omeprazole, which was
`launched in 2000.
`
`GASTROESOPHAGEAL REFLUX
`DISEASE
`Any symptomatic clinical
`condition with or without
`change in tissue structure that
`results from the reflux of gastric
`acid into the esophagus.
`
`HEARTBURN
`A burning sensation starting in
`the upper part of the abdomen
`and moving through the chest
`towards the throat.
`
`PEPTIC ULCERS
`Ulcers in the upper
`gastrointestinal tract, in which
`gastric acid is a key promoter.
`
`AstraZeneca R&D,
`431 83 Mölndal, Sweden.
`Correspondence to L.O.
`e-mail:
`lars.olbe@astrazeneca.com
`doi: 10.1038/nrd1010
`
`Gastric acid has been known for many decades to be a
`key factor in normal upper gastrointestinal functions,
`including protein digestion and calcium and iron
`absorption, as well as providing some protection against
`bacterial infections. However, inappropriate levels of
`gastric acid underlie several widespread pathological
`conditions, including GASTROESOPHAGEAL REFLUX DISEASE
`(GERD), for which HEARTBURN is the most common
`symptom, and PEPTIC ULCERS, which cause pain and suffer-
`ing in millions of people, and which, only thirty years
`ago, could be life-threatening if untreated. Treatment
`options then, however, were limited. For example, for
`peptic ulcers, the main treatment was administration
`of antacids to neutralize excess gastric acid (which
`promotes ulcer formation and prevents healing), but
`this provided only temporary relief. The alternative was
`an operation (gastrectomy, in which part of the stomach
`is removed, and/or vagotomy, in which nerves to the
`stomach are sectioned). The surgery could, however,
`have serious side effects. Pharmacological control of the
`complex mechanism of gastric acid secretion has there-
`fore long been desirable.
`The medical treatment of acid-related diseases — in
`particular peptic ulcers and GERD — had a break-
`through in the late 1970s with the introduction of the
`
`antisecretory drug cimetidine, an antagonist of the hista-
`mine 2 (H2) receptor, which has a key role in one of the
`pathways leading to gastric acid secretion. Cimetidine,
`and later comparable compounds with the same mecha-
`nism of action, have a marked gastric acid inhibitory
`effect, and considerably improved the lives of millions of
`people, as well as reducing the need for surgery.
`However, H2-receptor antagonists have a relatively short
`duration of action.
`From the late 1960s onwards, the pharmaceutical
`company Astra was also pursuing a programme
`aimed at finding a drug to inhibit acid secretion. In
`the 1970s, this led to the development of specific
`inhibitors of the proton pump in the acid-secreting
`parietal cells of the stomach, activation of which is
`now known to be the final step in acid secretion.
`These compounds were shown to be very potent
`inhibitors of gastric acid secretion, and demonstrated
`a surprisingly long-lasting duration of action. Ome-
`prazole — the first proton-pump inhibitor used in
`clinical practice — was launched in 1988 as Losec in
`Europe, and in 1990 as Prilosec in the United States.
`Omeprazole introduced a new approach for the effec-
`tive inhibition of acid secretion and the treatment of
`acid-related diseases, and was quite quickly shown to
`
`132 | FEBRUARY 2003 | VOLUME 2
`
`www.nature.com/reviews/drugdisc
`
`Page 1 of 8
`
`Patent Owner Ex. 2051
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`be clinically superior to the H2-receptor antagonists.
`None of the subsequently developed proton-pump
`inhibitors based on the omeprazole structure (but
`outside the original chemical patents) introduced by
`other companies have been shown to be significantly
`superior to omeprazole in clinical practice.
`During the 1990s, Astra tested several hundreds of
`compounds chemically based on the parent compound
`of omeprazole in order to find a proton-pump
`inhibitor with properties superior to omeprazole.
`Finally, esomeprazole emerged. Omeprazole is a race-
`mate consisting of two optical isomers (enantiomers),
`one being the mirror image of the other. The S isomer
`— esomeprazole — subsequently proved to be the first
`drug that is signficantly superior to omeprazole both as
`a gastric-acid inhibitor and for the clinical manage-
`ment of GERD. As predicted, the cause of the superior-
`ity of esomeprazole was higher bioavailability, which
`resulted in higher plasma concentrations than achiev-
`able with the R isomer. At the parietal-cell level, both
`isomers are equally effective, as both are transformed to
`the same active inhibitor within the parietal cell.
`Esomeprazole was launched as Nexium in 2000 by
`AstraZeneca. In this article, we summarize the develop-
`ment of omeprazole, focusing on the key discoveries
`and challenges, and then describe the subsequent
`development of esomeprazole (a more detailed history
`of the development of omeprazole can be found in the
`book Proton Pump Inhibitors1; see also REF. 2).
`
`Background
`In the late 1960s, the pharmaceutical company Hässle
`(a research company within Astra) decided to start a
`gastrointestinal research division with the aim of finding
`a potent drug for the inhibition of gastric acid secretion
`to be used in patients with peptic ulcers. To this end, a
`gastrointestinal laboratory was created, and the first
`project in this laboratory resulted in an antisecretory
`compound that was very effective in the rat, which was
`used as a screening model. However, the compound was
`completely ineffective in man, indicating that new
`screening models were needed.
`
`The omeprazole project
`In 1972, the gastric acid inhibitory project was restarted
`with a new approach. Anesthetized dogs were used as an
`initial screening model, followed by tests on conscious
`GASTRIC FISTULA DOGS. A literature search found a paper
`describing an antisecretory compound (CMN 131)
`developed by the pharmaceutical company Servier3; this
`compound, however, showed severe acute toxicity, and
`further research into this compound was consequently
`cancelled. As it seemed a reasonable assumption that the
`thioamide group in the chemical structure of CMN 131
`(FIG. 1) was responsible for the toxicity, the new approach
`aimed to eliminate this group by incorporating it into,
`or in between, heterocyclic ring systems. By 1973, the
`first hit was discovered — the benzimidazole H 124/26
`(FIG. 1), which was a powerful antisecretory compound
`without acute toxicity, and which therefore became the
`lead compound.
`
`R E V I E W S
`
`Patent problem. After H 124/26 had been identified, it
`was discovered that it was already covered by a patent
`owned by an Hungarian company, which described the
`compound as a drug for the treatment of tuberculosis.
`However, a metabolite of H 124/26, which was not
`included in the Hungarian patent, was found to be an
`even more potent antisecretory compound4. The
`metabolite H 83/69 was the sulphoxide of H 124/26 —
`named timoprazole (FIG. 1) — and it became the new lead
`compound. At this stage, the site of inhibitory action in
`the pathway leading to acid secretion was not known.
`
`Toxicological challenges. Long-term toxicological stud-
`ies of timoprazole revealed that it caused enlargement
`of the thyroid gland — later shown to be due to inhibi-
`tion of iodine uptake — as well as atrophy of the thymus
`gland. Thiourea compounds are well-known inhibitors
`of iodine uptake in the thyroid. A literature search of
`the chemistry of thiourea compounds showed a few
`substituted mercapto-benzimidazoles having no effect
`on iodine uptake, and the introduction of these sub-
`stituents into timoprazole resulted in elimination of the
`effects on the thyroid and thymus, without reducing the
`antisecretory effect. Tests on several substituted benzi-
`midazoles showed that separation of the inhibition of
`acid secretion from the inhibition of iodine uptake
`was obtained in a specific range of lipophilicity of
`these compounds5. The most potent antisecretory com-
`pound without thyroid/thymus effects was H 149/94,
`which was named picoprazole (FIG. 1).
`However, in extended toxicological studies of pico-
`prazole, as well as one previous compound, a few treated
`dogs developed NECROTIZING VASCULITIS. Fortunately from
`the perspective of the project, one of the control dogs
`also developed necrotizing vasculitis. It was shown that
`all the dogs with vasculitis emanated from one male
`dog, and all had antibodies against intestinal worms,
`which were probably obtained after deworming. New
`toxicological studies in another beagle strain, and in
`non-parasitized dogs in another laboratory, were com-
`pletely clean. Picoprazole was used in a concept study in
`human volunteers, and showed a potent antisecretory
`action of very long duration6.
`
`Compound optimization. Simpler in vitro techniques
`were essential in order to test a large number of different
`substituted benzimidazoles for the optimal inhibition of
`gastric acid secretion. The isolated gastric-acid-secreting
`mucosa of the guinea pig was introduced as an appro-
`priate in vitro model7. Later on, isolated rabbit acid-
`secreting glands were used8, and a micromethod for
`isolating acid-secreting glands from human gastric
`biopsies was developed9. These techniques allowed the
`testing of a large number of compounds, including tests
`on the human target tissue.
`At about this time, evidence was emerging that the
`activation of a newly discovered proton pump (an
`H+K+-ATPase) in the secretory membranes of the parietal
`cell was the final step in acid secretion10,11. Immuno-
`histological data obtained using antibodies against a
`crude preparation from the secretory membranes of
`
`GASTRIC FISTULA DOGS
`Dogs provided with a cannula
`into the stomach or into
`separated pouches of the
`stomach.
`
`NECROTIZING VASCULITIS
`An immunologically induced
`process causing an
`inflammatory reaction and
`necrosis in blood vessels.
`
`NATURE REVIEWS | DRUG DISCOVERY
`
`VOLUME 2 | FEBRUARY 2003 | 1 3 3
`
`Page 2 of 8
`
`Patent Owner Ex. 2051
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`Omeprazole was found to be the most potent
`inhibitor of stimulated gastric acid secretion in rats and
`dogs in vivo15, and no effects on iodine uptake, no induc-
`tion of thymus atrophy, no necrotizing vasculitis and no
`other signs of toxicity were found in initial safety studies.
`An Investigational New Drug (IND) application was filed
`in 1980, and omeprazole was taken into human trials in
`1982, which had highly encouraging results16–18. However,
`there were still further challenges to address.
`
`Further toxicological problems. Lifelong toxicological
`studies of very high doses of omeprazole in rats revealed
`the development of endocrine tumours (carcinoids)
`in the stomach, which led to the halting of all clinical
`studies in 1984. The carcinoids originated from entero-
`chromaffine-like (ECL) cells, a type of endocrine cell in
`the gastric mucosa that synthesize and secrete histamine
`in response to stimulation by the gastric hormone
`gastrin. However, longer-term stimulation by gastrin
`has a potent trophic action on ECL cells. Combining
`this with the fact that gastrin was known to be released
`in increasing amounts from the antrum of the stomach
`as the amount of acid secretion decreases suggested a
`possible explanation for the observed effects of lifelong
`very high doses of omeprazole in rats: the elimination of
`gastric acid secretion, resulting in massive hypergas-
`trinemia. This was shown to be the cause of the ECL cell
`hyperplasia in omeprazole-treated rats, as the hyper-
`plasia did not occur in rats subjected to resection of the
`gastric antrum19. Furthermore, the ECL cell carcinoids
`were also produced in lifelong studies of rats adminis-
`tered a H2-receptor antagonist (ranitidine) in high
`doses20, as well as by a surgical procedure that created
`massive hypergastrinemia21. These data allowed clinical
`studies with omeprazole to be restarted.
`
`Resumption of clinical studies. Omeprazole was found
`to be significantly superior to previous treatment
`regimens of H2-receptor antagonists in patients with
`duodenal17,18,22 and gastric ulcers23. A particularly
`notable superiority of omeprazole compared with the
`H2-receptor antagonist ranitidine was found in GERD
`patients24–26, in which the healing rates were about twice
`as high with omeprazole. On the basis of these studies,
`omeprazole was launched in Europe as Losec in 1988.
`Clinical doses of omeprazole produce a modest
`hypergastrinemia in the same range as the surgical
`procedure vagotomy27, and neither treatment has pro-
`duced ECL cell carcinoids over long-term (that is,
`greater than 10 years) follow-up. Massive hypergastrine-
`mia in man is seen in patients with gastrin-producing
`tumours, and these patients develop hyperplasia of
`the ECL cells, but not ECL cell carcinoids. Obviously,
`the response of the ECL cells to hypergastrinemia is
`different in man and rat.
`
`Mechanism of action of omeprazole. The success of
`omeprazole in the clinic can be ascribed to the very
`effective inhibition of gastric acid secretion achieved
`through specific inhibition of the gastric H+K+-ATPase.
`This proton pump is located in the secretory membranes
`
`NN
`
`H
`
`N
`
`S
`
`[H 124/26] (1973)
`
`CO2CH3
`
`CH3
`
`NN
`
`H
`
`CH3
`
`SO
`
`N
`
`OCH3
`
`Picoprazole
`[H 149/94] (1976)
`
`N
`
`N
`
`H
`
`CH3
`
`SO
`
`• •
`
`OCH3
`
`N
`
`Esomeprazole
`(1983)
`
`OCH3
`
`H3C
`
`R E V I E W S
`
`N
`
`S
`
`NH2
`
`[CMN 131]
`
`NN
`
`H
`
`SO
`
`N
`
`Timoprazole
`[H 83/69] (1974)
`
`NN
`
`H
`
`CH3
`
`SO
`
`OCH3
`
`N
`
`H3C
`
`Omeprazole
`[H 168/68] (1979)
`
`Figure 1 | Chemical milestones in the development of proton-pump inhibitors and the
`year of synthesis.
`
`parietal cells revealed strong immunoreactivity in the
`parietal cell region of the stomach, but also some
`activity in the thyroid gland12. Coupled with knowledge
`of the side effects of timoprazole on the thyroid dis-
`cussed earlier, these findings raised the intriguing possi-
`bility that benzimidazoles such as timoprazole could be
`inhibitors of the H+K+-ATPase. Research was initiated in
`this area in parallel with the further development of
`benzimidazoles, and it was indeed subsequently shown
`that substituted benzimidazoles inhibit gastric acid
`secretion by blocking the H+K+-ATPase13,14. The mode
`of action of substituted benzimidazoles, and the impli-
`cations of this for their clinical benefits, are discussed
`further in a following section.
`How could the antisecretory effect of substituted
`benzimidazoles be optimized? As weak bases accumu-
`late in the acidic compartment of the parietal cell close
`to the proton pump, substituents were added to the
`pyridine ring of timoprazole to obtain a higher pKa
`value, thereby maximizing the accumulation within
`the parietal cell. The result was compound H 168/68,
`which was named omeprazole (FIG. 1). It was later
`shown in a thorough mechanistic investigation that
`the higher pKa value of the omeprazole pyridine ring
`(~1 pKa unit higher than that in timoprazole) also
`increased the rate of acid-mediated conversion to the
`active species (the sulphenamide; see the section
`below on mechanism of action), which is the major
`factor determining acid inhibitory activity5. Also, the
`5-methoxy substitution pattern in the benzimidazole
`moiety of omeprazole made the compound much
`more stable to conversion at neutral pH compared with,
`for example, the ester substitution in the benzimidazole
`moiety of picoprazole (FIG. 1).
`
`134 | FEBRUARY 2003 | VOLUME 2
`
`www.nature.com/reviews/drugdisc
`
`Page 3 of 8
`
`Patent Owner Ex. 2051
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`R E V I E W S
`
`of the parietal cell of the gastric mucosa and constitutes
`the final step of acid secretion28 (FIG. 2a). Therefore,
`blockade of this pump results in a more specific inhi-
`bition of acid secretion compared with blockade of
`the more widely distributed H2 and cholinergic recep-
`tors. Furthermore, as omeprazole interacts with the
`final step of acid production, the inhibition of gastric
`
`acid secretion is independent of how acid secretion is
`stimulated29,30 — an important advantage over other
`pharmacological approaches to inhibiting acid secre-
`tion. For example, the inhibition of acid secretion by
`H2-receptor antagonists can be overcome by food-
`induced stimulation of acid secretion via gastrin or
`cholinergic receptors.
`
`a
`
`Histamine H2
`receptor antagonists
`
`Muscarinic
`M3 receptor
`
`Acetylcholine
`Muscarinic
`antagonists
`
`Histamine
`
`Histamine H2 receptor
`
`Gastrin
`
`CCK2 receptor
`
`cAMP-
`dependent
`pathway
`
`Ca2+-
`dependent
`pathway
`
`+
`
`+
`
`Parietal
`cell
`
`H+
`
`K+
`
`H+K+-ATPase
`
`Proton pump
`inhibitors
`
`Cl –
`
`Acid (HCl)
`
`Gastric gland
`lumen
`
`b
`
`Blood
`
`Parietal cell canaliculus lumen
`
`OCH3
`
`OCH3
`
`OCH3
`
`H+K+-ATPase
`
`OCH3
`
`H3C
`
`CH3
`
`H3C
`
`CH3
`
`N
`
`N
`
`H
`
`O
`
`S
`
`CH3
`
`H+
`
`N
`
`N
`
`H
`
`O
`
`S
`
`CH3
`
`3 steps
`
`N
`
`S
`
`N
`
`N
`
`SH
`
`N
`
`N
`
`S
`
`NH
`
`S
`
`H+K+-ATPase
`
`CH O3
`
`N
`
`CH O3
`
`NH+
`
`H3C
`Omeprazole
`
`H3C
`
`OCH3
`
`OCH3
`
`Sulphenamide intermediate
`
`Enzyme–inhibitor complex
`
`Figure 2 | Proton-pump inhibition. a | Gastric acid is secreted by parietal cells of the stomach in response to stimuli such as the
`presence of food in the stomach or intestine and the taste, smell, sight or thought of food. Such stimuli result in the activation of
`histamine, acetylcholine or gastrin receptors (the H2, M3 and CCK2 receptors, respectively) located in the basolateral membrane of
`the parietal cell, which initiates signal transduction pathways that converge on the activation of the H+K+-ATPase — the final step of
`acid secretion. Inhibition of this proton pump has the advantage that it will reduce acid secretion independently of how secretion is
`stimulated, in contrast to other pharmacological approaches to the regulation of acid secretion; for example, the inhibition of acid
`secretion by H2 receptor antagonists can be overcome by food-induced stimulation of acid secretion via gastrin or acetylcholine
`receptors. b | Proton-pump inhibitors such as omeprazole are prodrugs that are converted to their active form in acidic
`environments. Omeprazole is a weak base, and so specifically concentrates in the acidic secretory canaliculi of the parietal cell,
`where it is activated by a proton-catalysed process to generate a sulphenamide29. The sulphenamide interacts covalently with the
`sulphydryl groups of cysteine residues in the extracellular domain of the H+K+-ATPase — in particular Cys 813 — thereby inhibiting
`its activity30. The specific concentration of proton-pump inhibitors such as omeprazole in the secretory canaliculi of the parietal cell
`is reflected in their favourable side-effect profile.
`
`NATURE REVIEWS | DRUG DISCOVERY
`
`VOLUME 2 | FEBRUARY 2003 | 1 3 5
`
`Page 4 of 8
`
`Patent Owner Ex. 2051
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`R E V I E W S
`
`OCH3
`
`OCH3
`
`NN
`
`H
`
`OCH3
`
`N
`
`CH3
`O
`S
`
`Omeprazole
`
`Non-chiral oxidation
`
`H3C
`
`Mg(OCH3)2
`
`H3C
`
`OCH3
`
`N
`
`CH3
`O
`
`S
`
`N
`
`N
`1/2 Mg
`
`Esomeprazole magnesium
`(100% ee)
`
`OCH3
`
`NN
`
`H
`
`Asymmetric oxidation
`(Ti-mediated)
`
`OCH3
`
`H3C
`
`CH3
`
`N
`
`S
`
`H3C
`
`OCH3
`
`N
`
`CH3
`O
`
`S
`
`N
`
`N
`
`H
`
`OCH3
`
`Esomeprazole
`(94% ee)
`
`Figure 3 | Synthesis of omeprazole and esomeprazole. The large-scale production of esomeprazole is achieved by asymmetric
`oxidation of the same sulphide intermediate as is used in the production of omeprazole, which gives a 94% enantiomeric excess
`(ee). This is increased to 100% by preparing a magnesium salt of of esomeprazole and then performing a crystallization.
`
`So, how does omeprazole inhibit the H+K+-ATPase?
`In whole-body autoradiographic studies in mice,
`omeprazole was found to label only the tubulovesicles
`and secretory membranes of the parietal cell, which
`contain the H+K+-ATPase5. Electrophoretic analyses of
`such membranes, purified after administration of radio-
`labelled omeprazole, demonstrated that the radiolabel
`specifically associated with the 92-kDa proteins known
`to hold the catalytic subunit of H+K+-ATPase5. From
`this, it could be concluded that omeprazole binds only
`to the H+K+-ATPase in the gastric mucosa and nowhere
`else in the body.
`However, omeprazole itself is not the active inhibitor
`of the H+K+-ATPase. The transformation of omeprazole
`in acid is required to inhibit the H+K+-ATPase (FIG. 2b)
`in vitro and in vivo, whereas intact omeprazole is devoid
`of inhibitory action. Isolated H+K+-ATPase is blocked by
`omeprazole only after pretreatment of omeprazole with
`acid, and neutralization of the acidic secretory canaliculi
`of isolated gastric gland and parietal cell preparations by
`permeable buffers, which blocks the acid-catalysed
`transformation of omeprazole, prevents the inhibition of
`acid secretion. Furthermore, in vivo blockade of acid
`secretion using an H2-receptor antagonist prior to
`omeprazole administration decreases the inhibitory
`action of omeprazole. Investigations of the acid decom-
`position of omeprazole have revealed an intermediate
`compound — a sulphenamide — that effectively inhibits
`the H+K+-ATPase preparation in vitro 31 and which reacts
`rapidly with mercaptans (for example, β-mercap-
`toethanol) to form a disulphide adduct. As the H+K+-
`ATPase inhibition is associated with the modification of
`mercapto groups in the enzyme, such disulphide adducts
`can be considered as models of the enzyme–inhibitor
`complex, and the sulphenamide formed from omepra-
`zole can be considered to be the active inhibitor, which
`binds covalently to the cysteine residues (in particular,
`Cys 813) of the H+K+-ATPase (FIG. 2b).
`
`Omeprazole has several characteristics that are
`important for its unique mechanism of action. First,
`omeprazole is lipophilic, which means that it easily
`penetrates cell membranes. Second, it is a weak base,
`which means that it concentrates in acid compartments.
`Third, it is very unstable in an acidic solution. The half-
`life of omeprazole at pH 1 is ~2 minutes, whereas at
`pH 7.4 it is ~20 hours. So, omeprazole is a prodrug that
`accumulates within the acid space of the target cell,
`where it is transformed to the active inhibitor.
`Whereas the half-life of omeprazole in blood
`plasma is rather short — 1–2 hours in man — the
`half-life of the inhibitory complex is much longer. On
`the basis of the duration of action in humans, the
`half-life at the site of action is estimated to be ~24
`hours. Dissociation of the enzyme–inhibitor complex
`is probably a result of the effect of endogenous
`glutathione32,33, which leads to reactivation of the
`enzyme and the release of the omeprazole sulphide.
`The fact that the sulphide is found in the gastric juice
`is consistent with this idea. Reactivation of the acid-
`producing capacity may also in part be due to de novo
`synthesis of enzyme molecules34.
`
`Esomeprazole (Nexium)
`Omeprazole — need for improvement? Omeprazole
`showed a significant inter-individual variability, both
`regarding its pharmacokinetics and effect on acid secre-
`tion, and a significant number of patients with acid-
`related disorders needed higher or multiple doses to
`achieve symptom relief and healing. This difference in
`response was especially pronounced between slow and
`rapid metabolizers.
`In western countries, ~2–4% of individuals lack one
`of the isoenzymes — 2C19 — of the P450 enzyme
`family in the liver35. This isoenzyme is important for
`the metabolism of a number of drugs, including
`omeprazole35. Individuals lacking this isoenzyme
`
`136 | FEBRUARY 2003 | VOLUME 2
`
`www.nature.com/reviews/drugdisc
`
`Page 5 of 8
`
`Patent Owner Ex. 2051
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`R E V I E W S
`
`metabolize these drugs at a slower rate and are therefore
`called slow metabolizers. Among people from South
`East Asia and Japan, up to 20% are slow metabolizers35.
`On the basis of this knowledge, Astra started a new
`acid inhibitor research program in 1987 with the aim of
`finding a compound with reduced clearance by the
`liver — that is, increased bioavailability. Chemical
`approaches were used to change the substitution pattern
`on the pyridine and benzimidazole rings. More than 30
`scientists synthesized and screened several hundred com-
`pounds in the search for one that could possibly surpass
`omeprazole. During 1989–1994, four compounds passed
`the hard preclinical tests and demonstrated superior
`bioavailability compared with omeprazole in rats, and
`were then tested in humans. When all the key para-
`meters — pharmacokinetic properties, acid inhibitory
`effect and safety issues — were assessed, only one com-
`pound exceeded omeprazole: it was one of its optical
`isomers, the S isomer or esomeprazole.
`
`Optical isomers and consequences in biology. Around
`150 years ago, Louis Pasteur found that a solution of
`deposits from old wine casks could either rotate a
`beam of plane-polarized light to the right, to the left
`or not at all. He also found that an organic acid iso-
`lated from this tartar crystallized in two forms, and
`that a solution of one crystal form rotated the light to
`the left and the other form to the right, whereas a mix-
`ture of equal amounts of the two forms did not rotate
`the beam of light at all. Pasteur had detected iso-
`merism and is, therefore, the pioneer of stereochem-
`istry. Chemical compounds containing an atom
`(usually sulphur or carbon) bound to four different
`groups can occur in two forms, each of which is a
`non-superimposable mirror image of the other.
`Except for this ability to rotate plane-polarized light,
`their chemical structures and physicochemical prop-
`erties are the same36.
`In biology, stereochemistry is very important. Drug
`targets, such as enzymes and ion channels, as well as their
`endogenous ligands, hormones and signal substances,
`
`are the results of stereoselective biosynthesis. This
`means that drug targets recognize drug isomers and
`that the two isomers of a racemate often differ in their
`potency owing to different affinities for the target
`receptor, have different pharmacokinetic properties
`owing to different affinity for metabolizing enzymes or
`have different toxicological properties. Against this
`background and progress in chemical technology37,38,
`which has made it possible to synthetically produce
`pure isomers on a large scale, the regulatory authori-
`ties now require the development of a pure isomer
`whenever possible.
`
`Omeprazole is a racemate. The chemical structure of
`omeprazole contains a sulphoxide group (FIG. 1) and is
`therefore a racemate composed of the two isomers S
`and R in the proportion 1:1. As the two isomers of
`omeprazole have the same physicochemical proper-
`ties, they both undergo a non-enzymatic, proton-
`catalysed transformation to the active molecular
`species — the non-chiral sulphenamide (FIG. 2b). In
`accordance with this, we also found that the two iso-
`mers showed identical dose–response curves when
`tested in vitro for the inhibition of acid production in
`isolated gastric glands39.
`At this stage, however, we could not dismiss the idea
`of a possible difference in metabolism between the two
`isomers, but we needed larger amounts of the pure iso-
`mers. Isomer-selective production using microbial and
`enzymatic systems were only partly successful. We
`became more focused on the new idea of separating
`isomers via chromatography of diastereomers40. As a
`spin-off from our previous work, we identified a
`technique to use a temporary covalent diastereomeric
`complex with mandelic acid for chromatographic
`separation. Not only did we obtain hundreds of
`milligrams of the single isomers, but we also realized
`that alkaline salts, as opposed to the neutral forms,
`were crystalline and, moreover, that they were stable
`against racemisation. We now had the prerequisites
`for in vivo testing.
`
`90.7%
`
`64.5%
`
`25.3%
`
`b
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`Inhibition (%)
`
`R-omeprazole
`Omeprazole
`S-omeprazole (esomeprazole)
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`R-omeprazole
`
`Omeprazole
`
`Time after dose (hours)
`
`S-omeprazole
`(esomeprazole)
`
`1,500
`
`1,250
`
`1,000
`
`750
`
`500
`
`250
`
`a
`
`Mean plasma concentration (nmol/L)
`
`0
`
`0
`
`Figure 4 | Effects of racemic omeprazole and its enantiomers. a | Drug plasma concentrations and b | inhibition of
`pentagastrin-stimulated gastric acid secretion in healthy subjects (n = 4) after oral administration of 15 mg of R-omeprazole,
`omeprazole and esomeprazole at time 0 (REF. 42).
`
`NATURE REVIEWS | DRUG DISCOVERY
`
`VOLUME 2 | FEBRUARY 2003 | 1 3 7
`
`Page 6 of 8
`
`Patent Owner Ex. 2051
`Mylan v. Pozen
`IPR2017-01995
`
`

`

`The large-scale production of esomeprazole is now
`successfully achieved by asymmetric oxidation of the
`same sulphide intermediate as is used in the production
`of omeprazole (FIG. 3). Drawing inspiration from the
`work on titanium-mediated asymmetric oxidation
`reported by K. Barry Sharpless, who won the Nobel
`Prize for Chemistry in 2001, and H.B. Kagan, and on the
`basis of some key changes in the process parameters, the
`ENANTIOMERIC EXCESS could be increased from 4% to 94%.
`In the production method, the optical purity is further
`enhanced by the preparation of esomeprazole magne-
`sium salt with subsequent crystallization41.
`
`Isomer pharmacokinetics and pharmacodynamics.
`Initial in vivo experiments were carried out with the
`pure isomers and the racemates in rats and dogs to
`assess the the plasma concentration following oral
`doses and the effects on stimulated acid secretion. In
`the rat, the R isomer showed significantly higher
`bioavailability and more potent inhibition of acid
`secretion than the S isomer and the racemate. In the
`dog, however, no differences could be detected between
`the two isomers. If the initial in vivo experiments had
`been performed solely in dogs, we would have probably
`stopped further work with the isomers. On the basis of
`the initial findings in rats, however, we took the deci-
`sion to continue the project and compare the isomers
`in man. We found about the same magnitude of differ-
`ence between the isomers as in the rat, but, to our sur-
`prise, the S isomer had the highest bioavailability and
`oral potency in inhibiting gastric acid secretion in
`man42 (FIG. 4), owing to stereoselective metabolism of
`omeprazole43.
`In man, 40 mg of esomeprazole taken orally
`showed much higher and more prolonged plasma
`concentration curves than 20 mg esomeprazole (FIG. 5).
`The 40-mg esomeprazole dose also showed significantly
`
`Esomeprazole 20 mg
`Esomeprazole 40 mg
`Omeprazole 20 mg
`
`R E V I E W S
`
`ENANTIOMERIC EXCESS
`If one enantiomer is present
`to a greater extent, an
`enantiomeric excess exists
`where: enantiomeric excess =
`(measured specific rotation of
`mixture/specific rotation for the
`pure enantiomer) × 100.
`
`4,000
`
`3,000
`
`2,000
`
`1,000
`
`Mean plasma concentration (nmol/L)
`
`0
`
`0
`
`more potent inhibition of gastric acid secretion than
`the 20 mg esomeprazole, 20 and 40 mg omeprazole as
`well as the standard doses of the other commercially
`available proton-pump inhibitors44. Consequently,
`40 mg of esomeprazole was chosen as the standard
`dose when esomeprazole was launched as Nexium in
`2000. Reassurance on the safety of a high standard
`dose of esomeprazole was provided by the extensive
`experience of continuous treatment with both 20 mg
`and 40 mg of omeprazole45, in contrast to the situation
`in which the choice of standard 20 mg dose of
`omeprazole was made. At that time, the safety of the
`drug was more uncertain, as it was a completely new
`type of drug, and the choice of dose had to be the
`lowest possible that still retained sufficient activity.
`
`Clinical results. In GERD, the most common acid-
`related disease, esomeprazole has been shown to be
`clinically superior to other proton-pump inhibitors,
`such as omeprazole and lansoprazole. There is signifi-
`cantly higher symptom relief during treatment with
`esomeprazole46,47, and significantly higher healing
`rates, which are well above 90% in patients with eso-
`phagitis46,48,49. The remission rate during maintenance
`treatment is also significantly higher with esomeprazole
`in comparison with placebo and lansoprazole50,51.
`It is now known that most peptic ulcers are caused
`by infection with Helicobacter pylori, with gastric acid
`still being a key promoter. Elimination of H. pylori is
`now the clinical approach of choice52. At present, anti-
`biotics alone are not sufficient to achieve high enough
`cure rates, and so two antibiotics are combined with an
`antisecretory agent, such as a proton-pump inhibitor,
`in ‘triple therapies’. For esomeprazole, one-week triple
`the