The purpose of this research study is to learn about whether treating the esophagus with amiloride reduces either the frequency or the time to onset of acid-induced heartburn in patients with nonerosive reflux disease. In particular, we are looking at people who have either had complete relief while using a Proton Pump Inhibitor (PPI) or who have only had some relief of symptoms while on a PPI.

It is now well established that patients with gastroesophageal reflux disease - both erosive and non-erosive forms - have on esophageal biopsy a lesion in esophageal stratified squamous epithelium known as 'dilated intercellular spaces (DIS). This lesion has importance because it reflects an increase in paracellular permeability due to acid damage to the tight junctions [1]. Moreover, the increase in paracellular permeability in non-eroded esophageal epithelium provides a plausible explanation for why heartburn occurs during episodes of acid reflux (or esophageal acid perfusion) in those with nonerosive reflux disease (NERD) but not in healthy subjects, and that is because luminal acid is now able to diffuse between cells in quantities sufficient to acidify the intercellular space [2]. Further, the lowering of intercellular pH is the likely trigger for heartburn by its ability to stimulate pain-sensing neurons (nociceptors) within the esophageal mucosa; with the signal from these neurons being transmitted to the CNS for heartburn perception. Clinical and experimental evidence support these concepts in that antacid ingestion relieves and proton pump inhibitors (PPIs) prevent heartburn in most patients and when PPIs control heartburn, they also lead to resolution of DIS in squamous epithelium [3]. Moreover, it is increasingly likely that many NERD patients classified as PPI-partial responders develop heartburn through the same mechanism as PPI-complete responders with NERD. This is because PPIs only raise gastric pH to ~pH 5, and even such 'weakly-acidic' refluxates have been shown to be associated with heartburn [4]. The reason for this is that in the presence of a broken epithelial barrier, weakly acidic refluxates are still able to acidify the intercellular space to levels sufficient to activate the nociceptors - the latter the case since the nociceptors can be activated even at pHs as modestly acidic as pH 6.0-pH 7.0 (see below).

The esophageal mucosa has two nociceptors that are candidates to mediate heartburn in NERD. These are: a) a capsaicin-sensitive, transient receptor potential vanilloid receptor (TRPV)-1, and b) an amiloride-inhibitable, acid-sensing ion channel (ASIC), subtype 2. Both nociceptor types innervate the esophageal mucosa and are activated by small declines in environmental (intercellular) pH [5,6]. Recently, a role for TRPV-1 was sought in the causation of heartburn during esophageal acid perfusion [7]. This was done by perfusing the esophagus with capsaicin in quantities presumed sufficient to desensitize TRPV-1 and then perfusing with acid to see if it blocked heartburn. The results, however, were negative in that acid still elicited heartburn. From this, one can conclude that heartburn is either not mediated by TRPV-1 or that capsaicin failed to adequately desensitize TRPV-1. We propose to test the hypothesis that capsaicin's failure to block heartburn in these subjects was because the actual nociceptors mediating heartburn are mucosal ASICs rather than TRPV-1. The hypothesis will be tested in a double-blind crossover study designed to determine if esophagal perfusion with the ASIC-inhibitor, amiloride, can block heartburn elicited by acid-perfusion in PPI-complete responders and PPI-partial responders with NERD [8]. The expectation is that, when compared to placebo, amiloride, which readily diffuses across intact esophageal epithelium, will reduce the frequency and both prolong the onset and reduce the severity of heartburn elicitable by esophageal acid perfusion. Such an outcome would provide support for mucosal ASICs, rather than TRPV-1, as mediator of heartburn in NERD, and raise interest in ASICs as a potential therapeutic target for that subset with NERD that are PPI-partial responders.

Inclusion Criteria:

• Adults, 18-59 years old
• moderate heartburn at least 3 days/week
• males and non-pregnant/non-lactating females
• Complete relief while using a PPI or only some relief of symptoms while on a PPI

Exclusion Criteria:

• erosive esophagitis
• unable or unwilling to undergo endoscopy and biopsy or Bernstein testing
• eosinophilic esophagitis
• negative Bernstein test
• known hypersensitivity to amiloride
• renal disease
• diabetes
• hypotension
• electrolyte imbalance
• contraindication to diuretics, including taking lithium or ACE inhibitors. -history of gastric or esophageal surgery
• history of ZE syndrome
• bleeding disorder
• UGI bleeding
• esophageal motor disorder
• esophageal stricture
• Barrett's esophagus
• UGI malignancy
• esophageal varices
• subjects with current malabsorption
• inflammatory bowel disease
• severe heart-lung-liver-renal-cerebrovascular disease
• subjects post-transplant
• diabetes
• actively taking the following medications: tricyclic antidepressants, quinidine, quinine, dilantin, warfarin, narcotic analgesics, antineoplastic agents, salicylates (except a baby aspirin for cardiovascular protection); steroids, NSAIDs (including COX-2 inhibitors), KCl, anti-tuberculosis medication, bisphosphonates, and triamterene, cyclosporine, tacrolimus, and other potassium sparing drugs like spironolactone
• serum potassium of 5.5 mEq/L or higher