Effects of YF476 and Rabeprazole on Gastric Function

We want to know if YF476 prevents ECL-cell hyperplasia induced by PPI, and rebound hyperacidity after PPI withdrawal, in man, as it does in the rat. A positive result would support further studies to assess if a combination of YF476 and a PPI is a better way of treating patients with peptic ulcer disease and GORD than a PPI alone. Equally, we want to know if YF476 by itself not only inhibits gastric acid secretion but also does not cause ECL-cell hyperplasia and rebound hyperacidity after withdrawal. A positive result would support further studies to assess if YF476 alone is an alternative and perhaps better treatment for patients with peptic ulcer disease and GORD than a PPI alone. The study design is important. We need to ask ourselves several questions.

First, which PPI should we use? Omeprazole has been studied more than other PPI, and is probably still the most widely used PPI in the clinic. A disadvantage of omeprazole is that its metabolism is affected by genetic polymorphism. Studies of human liver microsomes in vitro (Yamazaki et al 1997) show that omeprazole is metabolised by 5-hydroxylation (catalysed by CYP2C19, and to a lesser extent CYP3A4) and sulphoxidation (catalysed by CYP3A4). YF476 inhibits CYP3A4/5 in vitro; but we don't know how relevant that is in vivo. CYP2C19 has 3 genotypes: homozygous extensive metabolisers, with higher enzymatic activity; heterozygous extensive metabolisers, with moderate enzymatic activity; and poor metabolisers, with markedly impaired enzymatic activity. Consequently, in CYP2C19 homozygous extensive metabolisers, acid suppression by omeprazole is reduced compared with heterozygous and poor metabolisers (Shirai et al 2001). Homozygous extensive metabolisers comprise about 70% of the European and USA populations and about 30% of the Asian population. Recently, a new CYP2C19 gene variant that causes ultrarapid metabolism of omeprazole has been identified in 18% of Swedes (Sim et al 2006). So, if we use omeprazole, we should genotype subjects for CYP2C19. Ideally, we should stratify their allocation to treatments, to try to avoid confounding the results. But, we are unlikely to be able to get the genotyping results quickly enough after screening to make stratification practical.

Rabeprazole, unlike omeprazole, is metabolised mainly via a non-enzymatic pathway with minor CYP2C19 and CYP3A4 involvement. Therefore, acid suppression by rabeprazole is less affected by CYP2C19 genotype (Shirai et al 2001; Miura et al 2006), which is one reason for using rabeprazole instead of omeprazole. Omeprazole and rabeprazole are both off-patent, so a combination product of YF476 and either PPI is possible.

Second, what dose of PPI should we use, and for how long? The dose of PPI should be one that is clinically relevant. If the choice of PPI is omeprazole, the dose should be 40 mg daily, because 20 mg -- the standard clinical dose -- was less effective than the recommended doses of other PPI in raising gastric pH (Warrington et al 2006). Rabeprazole 20 mg was more effective than esomeprazole 20 mg in those studies, another reason for using rabeprazole (Warrington et al 2005). Taking everything into account, we have decided to use rabeprazole 20 mg daily for our study.

The duration of PPI dosing is probably more important than the choice of PPI and dose. The longer the duration of dosing, the more likely we are to show the morphological changes of ECL-cell hyperplasia. But ethical considerations and compliance issues mean that we have to compromise over duration of dosing. The available evidence suggests that 2 weeks is not long enough to induce ECL-cell hyperplasia, whereas 8-12 weeks may be more than long enough. So, 6 weeks seems a reasonable compromise.

Third, how long after stopping the PPI should we assess rebound hyperacidity? Early studies in man failed to show rebound hyperacidity after stopping omeprazole because the assessments were done too soon (Prewett et al 1991). PPIs are non-competitive inhibitors of the proton pump, which takes days to regenerate fully. The available evidence suggests that 2 weeks after PPI withdrawal is long enough to study rebound hyperacidity.

Fourth, what dose of YF476 should we use? Single and repeated doses of 100 mg daily were well tolerated and completely suppressed pentagastrin-stimulated increases in gastric acid volume and H+ content in healthy subjects. The toxicology and toxicokinetic studies support dosing of 100 mg daily for up to 13 weeks. So, a dosing schedule of 100 mg daily for 6 weeks is appropriate.

Fifth, is it safe to give a PPI and YF476 together? No safety problems were reported when rats were given omeprazole and YM022, an enantiomer of YF476, for 13 weeks (Nishida et al 1995). Also, no safety problems were reported when omeprazole and YF476 were given to rats for 8 weeks (Chen et al 2000). However, whatever PPI we use, we should assay blood concentrations of PPI and YF476 (Redrup et al 2002) at the start and end of dosing, lest a drug-drug interaction confound the results of the outcome measures.

Sixth, what tests of gastric function and ECL cells should we use to compare the treatments? We should study: plasma gastrin, to assess the degree of hypergastrinaemia; plasma CgA, a marker of ECL-cell hyperplasia; plasma SST, a marker of D-cell activity; pentagastrin-stimulated increases in gastric acid volume and H+ content, and 24-hour ambulatory pH, to compare anti-secretory responses during and after stopping treatments; gastric biopsies, to assess changes in ECL-cell histology and enzymes; and dyspepsia symptoms, to assess occurrence of symptomatic rebound hyperacidity after stopping treatment.

Seventh, should the study be crossover or parallel-group in design? In healthy subjects, a crossover study is often the ideal design. Comparisons are made within rather than between subjects, which reduces the variability of the results. But, a crossover design is not practical given the long dosage period and the demanding procedures. By using rabeprazole, we avoid the potential problem of between-subject differences attributable to CYP2C19 genetic polymorphism. So, a parallel-group design with 3 groups of subjects randomised to either YF476, rabeprazole, or a combination of YF476 and rabeprazole, is a sensible solution.

Eighth, should we have identical treatments and a placebo group? We can over-encapsulate the treatments to make them identical so that the study is double blind in design. Ideally, we should have a placebo group, but if the study is double blind in design and the methods are mostly objective, a placebo group is not essential.

Finally, should we use patients or healthy subjects? Healthy subjects and patients with peptic ulcer disease or GORD all develop hypergastrinaemia and ECL-cell hyperplasia after repeated dosing with a PPI, and rebound hyperacidity after stopping the PPI. Nevertheless, the study is best done in healthy subjects because they are more likely to comply with the demanding study procedures, more homogeneous, more robust, and not on medication. The YF476 reproductive toxicology studies justify including women who are not at risk of pregnancy. However, H pylori-positive subjects should be excluded. In a recent UK community study, 15% of people were H pylori-positive (Lane et al 2006).