Defective barrier function in neosquamous epithelium

Abstract

Objectives: Radiofrequency ablation (RFA) of Barrett's esophagus (BE) is a common strategy for the prevention of esophageal adenocarcinoma (EAC). After RFA, the ablated esophagus heals on acid suppressive therapy, and is re-populated with a stratified squamous epithelium, referred to as "neosquamous epithelium (NSE)." Because the ability of the NSE to protect the underlying tissue from recurrent insult by reflux is unclear, we assessed the barrier function of NSE by comparing it to that of the native upper squamous epithelium (USE) in subjects having undergone RFA.

Methods: At varying intervals following RFA, the barrier function of NSE and USE were assessed in endoscopic biopsies by light and electron microscopy, and by measurement of electrical resistance (R) and fluorescein flux in mini-Ussing chambers. Chamber results were further compared with results from control biopsies (healthy distal esophagus). A claudin expression profile in the tight junctions (TJs) of NSE and USE was determined using Quantitative reverse transcriptase PCR. Differential expression of claudin-4 between NSE and USE was assayed by immunoblots.

Results: USE was histologically normal whereas NSE showed dilated intercellular spaces and marked eosinophilia. NSE was also more permeable than USE and healthy controls, having lower mean R and higher fluorescein fluxes. Abnormally low R values for NSE were unrelated to the time period following RFA (or number of prior RFA sessions), being abnormal even 26 months after RFA. Abnormal permeability in NSE was associated with significantly lower values for claudin-4 and claudin-10 than in USE.

Conclusions: NSE commonly exhibits defective barrier function. As this defect will make it vulnerable to injury, inflammation, and destruction by acidic and weakly acidic refluxates, it may in part explain incidences of recurrence of BE following ablation.

Fig. 1

A light photomicrograph of a hematoxylin-eosin stained section of native esophageal upper squamous epithelium (left panel) and neosquamus epithelium from the lower esophagus (right panel). Note that neosquamous epithelium shows dilated intercellular spaces (white arrows) and has a prominent infiltrate of eosinophils (black arrows) when compared with the normal-appearing ‘native’ upper squamous epithelium. Magnification 60X. The INSERT in the right panel is an electron photomicrograph to better illustrate the dilated intercellular spaces in neosquamous epithelium. Magnification 3000X.

Fig. 2

The transepithelial electrical resistance (R_T) (upper panel) and fluorescein flux (lower panel) are illustrated for neosquamous epithelium (NSE), native esophageal upper squamous epithelium (USE) from the same patients, and healthy native distal squamous epithelium from subjects without esophageal disease (controls). Note that NSE has significantly lower R_T and higher fluorescein flux values than both USE and healthy controls indicating higher paracellular permeability to ions and uncharged molecules, respectively. * p<0.05 compared to healthy controls.

Fig. 3

A plot of the transepithelial electrical resistance (R_T in ohms. cm²) in NSE versus the time (months) from the last performance of radiofrequency ablation of Barrett’s esophagus. Note that there is no trend for R_T to improve with time for NSE (linear regression R²=0. 008; p=0.7).

Fig. 4

Claudin gene expression profiles for neosquamous epithelium (NSE) and for native esophageal upper squamous epithelium (USE). Claudin expression levels are referenced to expression levels for ZO-1 which is set at 1.0. Note that expression of claudin-4 and claudin-10 were significantly lower in NSE than in USE while there was no difference in expression of the most prominently expressed claudin, i.e. claudin-1. Error bars = [(2^%CV)/100] x [relative expression]; NT = not tested; *p<0.05 compared to USE.

Fig. 5

An immunoblot of claudin-4 for neosquamous epithelium (NSE) and for native esophageal upper squamous epithelium (USE). Ten micrograms of protein were loaded in each lane and the actin signal shown to document equal protein loading. Molecular weight standards are displayed on the left in kilodaltons (kDa). Prestained SDS-PAGE standard (wide range; Bio-Rad) was used (red bands). Claudin-4 is shown to be more strongly expressed in USE than NSE.