| OCR Text |
Show 35 2 | PETERSON 1 | BACKGROUND ET AL. consensus criteria2—requiring oesophageal eosinophilia on biopsy (≥15 eos/HPF) whose eosinophilia remained ≥15 eos/HPF after Although a close relationship between eosinophilic oesophagitis 8 weeks of double-dose proton pump inhibitor therapy. Chart review (EoE) and proton pump inhibitor-responsive oesophageal eosinophi- and other biopsies were also reviewed to confirm that other EoE lia (PPI-REE) has long been hypothesised,1 American guidelines for consensus criteria were met, such as the presence of appropriate eosinophilic oesophagitis have distinguished PPI-REE from EoE.2,3 oesophageal symptoms and lack of comorbid disease such as Crohn’s Since then, extensive evidence has shown that, other than by pro- disease, etc. PPI-REE patients were defined as those patients with ton pump inhibitor-response, these entities are mostly indistin- ≥15 eos/HPF on oesophageal biopsy prior to proton pump inhibitor guishable by clinical presentation,4-9 endoscopic appearance4-7,9 therapy whose eosinophilia fell to <10 eos/HPF after 8 weeks of and histology.4,5,7,9 The one exception is patients with more high-dose proton pump inhibitor therapy. Controls were age- and 4 10 oesophageal eosinophil infiltrates are gender-matched patients with no history of current or prior oeso- modestly less likely to be PPI-responsive. A recent European task phageal disease whose oesophageal biopsies were histologically nor- force concluded that PPI-REE is highly similar to EoE and poten- mal. We found 64 patients with oesophageal biopsies meeting these tially an EoE treatment variant.11 Their main qualm was that it criteria between 2012 and 2015. Oesophageal biopsies were located had not yet been shown that PPI-REE cases respond to dietary on 61 patients. Biopsies were examined for depth of tissue, eosino- exclusion and thus antigen-induced immune responses. Shortly phil counts, and gastric or other tissue contamination. Twenty sam- thereafter, 2 studies showed that proton pump inhibitor-responsive ples were excluded because they either had gastric, intestinal, or cases can also respond to food exclusion.12,13 In addition to glandular metaplastic tissue contamination or had extensive subep- reducing gastric acidity (which reduces acid-related effects includ- ithelial tissue (extending 10 lm or more beneath the epithelium). ing altered permeability), PPIs, via Signal Transducer and Activator The remaining tissues (14 EoE, 13 PPI-REE and 14 controls) were of Transcription 6 (STAT6), inhibit synthesis of eotaxin-3,14,15 the examined by RNA sequencing. Age, atopy, serum total IgE and gen- dominant EoE chemokine. der were recorded. florid or more extensive Prior microarray-based gene expression studies comparing PPIREE and EoE showed modest differences in the transcriptome, but these studies were either limited to 94 transcripts16 or studied a limited number of subjects (6 PPI-REE and 4 EoE).17 Wen et al16 found 2.1 | RNA isolation, sequencing and genome alignment that potassium voltage-gated channel subfamily J member 2 gene Total RNA was isolated from 10-lm-thick, formalin-fixed, paraffin- (KCNJ2) was a potential distinguishing factor in a study with 33 EoE embedded tissue using the High Pure FFPET RNA isolation kit (Roche, and 28 Pre-PPI-REE subjects. The other study failed to identify this Indianapolis, Indiana). DNA was removed from each RNA sample by transcript as 1 of 35 transcripts found to be differentially regu- the manufacturer’s suggested DNase 1-digestion step. RNA quantity lated.17 To our knowledge, neither study assessed MAPK8IP2 gene was determined with a NanoDrop spectrophotometer (Thermo Fisher (C-jun-amino-terminal kinase-interacting protein 2). Scientific, Waltham, MA, USA). Ribosomal RNA was removed from Proton pump inhibitor-responsive oesophageal eosinophilia is each RNA sample prior to cDNA library preparation using the Illumina worthy of study because it is a common finding. In prospective stud- TruSeq Stranded Total RNA Sample Prep Protocol. An Illumina HiSeq ies of patients presenting with oesophageal eosinophilia (≥15 eosino- 2500 instrument was used to sequence the polymerase chain reaction phils/maximal high power field [abbreviated as eos/HPF]), 36%-69% (PCR) amplified libraries using 50-cycle single-read chemistry. of patients have <15 eos/HPF after proton pump inhibitor ther- Sequencing reads were aligned to the GRCh37/Hg19 human refer- apy.4,9,18-21 RNA sequencing (RNA-Seq) quantifies a large dynamic ence genome using the Novoalign application (Novocraft, Selangor, range of gene expression levels, improving the ability to detect dif- Malaysia). The mean aligned read counts for the 41 formalin-fixed, ferential expression. We hypothesised that although PPI-REE is clo- paraffin-embedded RNA sequencing data sets was 14.9 million reads. sely related to EoE, there still could be transcriptome differences between them. We sought to further examine whether there are previously undetected differences between these 2 groups. 2.2 | Differential expression and clustering The USeq DefinedRegionDifferentialSeq (DRDS) application was used to count reads intersecting exons of each annotated gene and 2 | METHODS score them for differential expression using DESeq2 negative bino- To assess EoE and PPI-REE gene expression, under Institutional sequencing data were accepted as differentially expressed if they Review Board (University of Utah #63027) approval, we examined differed 2-fold or more with an adjusted p/false discovery rate 14 subjects with EoE by consensus criteria, 13 subjects with PPI- (FDR) of ≤0.01. Visualisation tracks were prepared for each sample REE and 14 controls. Each biopsy was done prior to any proton group using the USeq application Sam2USeq and viewed using the pump inhibitor therapy. EoE patients were defined using 2011 Integrated Genome Browser (IGB). Hierarchical clustering and mial statistics with Benjamini-Hochberg adjustments.22 RNA |
