Annals of Gastroenterology and Digestive Disorders (AGDD)
ISSN: 2643-2609 | Vol. 9, Issue 1, 2026, pp. 12-16

Eosinophil-Derived Transcript Signatures in Oesophageal Biopsies and Peripheral Blood from Patients with Eosinophilic Oesophagitis: A Review

Henrik Nilsson1, Marta Kowalczyk2, Pierre Lambert3

1Department of Gastroenterology, Lund University, Lund, Sweden
2Department of Paediatric Gastroenterology, Medical University of Warsaw, Warsaw, Poland
3Department of Mucosal Immunology, KU Leuven, Leuven, Belgium
DOI: | Published: 2026-08-17 | Article page and PDF

Abstract

Eosinophilic oesophagitis is diagnosed by counting eosinophils in a biopsy, which is a subjective measure of a patchy disease. Transcript profiling of the same tissue performs better than the count it is measured against. Blood, which would spare the endoscopy altogether, does not. Objectives: To review the published evidence on eosinophil-derived and epithelial transcript signatures in eosinophilic oesophagitis: what the oesophageal transcriptome contains, how a targeted molecular panel performs against histology, and how far peripheral blood markers track the tissue. Material and Methods: Narrative review of transcriptome and expression profiling studies, molecular panel validations, biomarker correlation cohorts, consensus criteria and histology scoring studies indexed in PubMed and the major gastroenterology and allergy journals. Studies were eligible if they reported the accuracy of a molecular or blood marker against oesophageal histology in a defined population, or the reproducibility of the histological reference itself. No new patient data were generated; every estimate quoted is one published by the original investigators. Results: Genome-wide profiling identified a transcript signature involving approximately one per cent of the human genome, conserved across sex, age

and allergic status and distinct from non-eosinophilic chronic oesophagitis, with the eosinophil chemoattractant eotaxin-3 the most highly induced transcript. A targeted 96-gene panel derived from that signature identified adult and paediatric patients with approximately 96 per cent sensitivity and 98 per cent specificity across 194 validation samples, retained 96 per cent sensitivity and 100 per cent specificity on archival formalin-fixed material, and distinguished disease in remission from controls. In peripheral blood, absolute eosinophil count, eosinophil-derived neurotoxin and eotaxin-3 correlated with oesophageal eosinophil density at r = 0.56, 0.54 and 0.32 respectively, while eotaxin-1, eotaxin-2, interleukin 5 and faecal neurotoxin did not correlate at all. Conclusion: The tissue transcriptome outperforms the eosinophil count that serves as its reference standard, and does so on the archival material most

laboratories already hold. Blood markers rise in active disease but correlate only moderately with tissue, which makes them plausible for monitoring a known patient and inadequate for diagnosis. Abbreviations: EoE – Eosinophilic Oesophagitis, EDP – EoE Diagnostic Panel, CCL26 – Chemokine (C-C Motif) Ligand 26 (Eotaxin-3), AEC – Absolute Eosinophil Count, EDN – Eosinophil-Derived Neurotoxin, eos/hpf – Eosinophils per High-Power Field, FFPE – Formalin-Fixed Paraffin-Embedded,qPCR – Quantitative Polymerase Chain Reaction, EoEHSS – EoE Histology Scoring System, PPI – Proton Pump Inhibitor, GORD – Gastro-Oesophageal Reflux Disease, IL – Interleukin.

Keywords: Eosinophilic Oesophagitis; Transcriptome; Eotaxin-3; Molecular Diagnosis; Blood Biomarkers

Introduction

Eosinophilic oesophagitis is a chronic, immune-mediated disease of the oesophagus whose recorded incidence has risen steadily across population-based studies [1,2]. It is defined clinically by symptoms of oesophageal dysfunction and histologically by at least fifteen eosinophils per high-power field, a threshold carried through successive consensus documents [3,4,5,6,7].

That definition has two structural weaknesses. The disease is patchy, so a biopsy from one site may not represent the oesophagus [8], and counting eosinophils is a subjective exercise with imperfect agreement between pathologists [9,10]. A scoring system incorporating additional features was developed partly for that reason and has been reported to outperform the peak count itself Molecular profiling offered a different route. Genome-wide analysis of oesophageal tissue identified a conserved disease-specific transcript signature [12], which was subsequently condensed into a targeted diagnostic panel [8,13]. The attraction is objectivity: a transcript measurement does not depend on who is looking down the microscope.

A separate line of work has asked whether the same information is available in blood, which would remove the endoscopy from diagnosis and monitoring altogether [14,15]. This review sets out what the tissue transcriptome achieves, what the blood markers achieve, and why the two answers differ.

Material and Methods

Search strategy: PubMed and the archives of the major gastroenterology and allergy journals were searched for studies of molecular and blood markers in eosinophilic oesophagitis. Search terms combined eosinophilic oesophagitis with transcriptome, gene expression, diagnostic panel, eotaxin-3, CCL26, absolute eosinophil count, eosinophil-derived neurotoxin, histology scoring and diagnostic accuracy. Reference lists of retrieved consensus documents and panel validations were screened by hand.

Eligibility: Studies were included if they reported the accuracy of a molecular panel or blood marker against oesophageal histology in a defined population, the composition of the disease transcriptome, or the reproducibility of histological assessment. Consensus documents were included where they define the reference standard in current use. Paediatric and adult cohorts are both included and identified as such.

Data extraction and handling: Design, sample size, tissue or fluid sampled, platform, comparator and reported accuracy or correlation estimates were extracted as published. Correlation coefficients are reported rather than converted, and the reference standard is stated alongside every accuracy figure, because in this field the reference is itself imperfect. No pooling or re-analysis was performed and no patient-level data were obtained. Figures 1 to 3 and Tables 1 to 3 reproduce published values only. This is a narrative synthesis and was not registered as a systematic review.

Results

The disease has a conserved transcript signature: Genome-wide microarray analysis of oesophageal tissue identified a signature involving roughly one per cent of the human genome. It was remarkably conserved across sex, age and allergic status, and was distinct from the profile of non-eosinophilic chronic oesophagitis. The most highly induced transcript was CCL26, encoding the eosinophil-selective chemoattractant eotaxin-3, whose messenger RNA and protein levels correlated strongly with tissue eosinophilia and mastocytosis [12].

That the signature separates eosinophilic from other chronic oesophagitis is the point that matters diagnostically, because reflux disease is the principal differential and eosinophil counts alone do not reliably distinguish the two [16,17]. Earlier work had established the underlying T helper type 2 pattern of the inflammatory response [18,19,20].

A targeted panel outperforms the count it is measured against:

The full transcriptome is impractical for routine use because of cost and turnaround, so a 96-gene quantitative polymerase chain reaction array with a dual clustering and dimensionality-reduction algorithm was derived from it. Developed on biopsies from 15 patients with disease and 14 controls, it was then applied to a separate cohort of including patients in remission, 34 histologically ambiguous samples and 12 with reflux [13].

The panel identified adult and paediatric patients with approximately 96 per cent sensitivity and 98 per cent specificity. It distinguished patients in remission from controls, identified patients exposed to swallowed corticosteroids, and separated disease from reflux oesophagitis confirmed by pH-impedance testing (Figure 1) [13]. In a substudy of 45 randomly selected archival formalin-fixed samples it retained 96 per cent sensitivity and 100 per cent specificity against histology, with signature acquisition apparently unaffected across three years of archiving [13,21].

Two further properties matter more than the headline accuracy. Preliminary evidence suggested the panel could identify patients likely to relapse after treatment [13,21], and a later analysis found that it detected inflammation remote from the biopsy site without loss of sensitivity, which directly addresses the patchiness that

Figure 1: Accuracy of the 96-gene panel against histology on fresh and archival tissue [13]
Figure 1: Accuracy of the 96-gene panel against histology on fresh and archival tissue [13]

Study n Material Principal reported finding

Straumann et al. 2001 Th2-type inflammatory — Tissue [18] response characterised Blanchard et al. 2006 Conserved signature; CCL26 — Tissue [12] most induced transcript Konikoff et al. 2006 AEC and EDN correlate with 47 Blood [14] tissue at r = 0.56 and 0.54 96-gene panel, 96% Wen et al. 2013 [13] 194 Tissue sensitivity and 98% specificity Histology score outperforms Collins et al. 2017 [9] — Tissue peak eosinophil count Consens Updated diagnostic criteria; Dellon et al. 2018 [3] — us PPI reclassified as treatment Table 1: Principal studies reviewed, with the material examined and the finding as reported by the original investigators What blood delivers, and what it does not: A prospective cross-sectional study of 47 paediatric patients undergoing endoscopy measured peripheral blood absolute eosinophil count, eosinophil-derived neurotoxin, eotaxin-1, -2 and -3, and interleukin 5, together with faecal neurotoxin, and correlated each with oesophageal eosinophil density [14].

Three markers correlated: absolute eosinophil count at r = 0.56 (p < 0.0001), neurotoxin at r = 0.54 (p < 0.0001) and eotaxin-3 at r = 0.32 (p = 0.04) (Figure 2). Eotaxin-1, eotaxin-2, interleukin 5 and faecal neurotoxin did not correlate with tissue eosinophilia and were not raised in active disease [14]. All three of the positive markers count 440 against 140 per microlitre, neurotoxin 50.3 against 31.1 nanograms per millilitre, and eotaxin-3 37.7 against 11.5 picograms per millilitre (Figure 3) [14].

The distinction to draw is between a marker that differs between groups and a marker that tracks an individual patient. A correlation of 0.56 explains under a third of the variance in tissue eosinophilia, so a blood result cannot substitute for the biopsy at diagnosis. The same figure is more defensible for following a patient whose diagnosis is already established, which is how the original

Figure 2: Correlation of blood markers with oesophageal eosinophil density in 47 patients [14]
Figure 2: Correlation of blood markers with oesophageal eosinophil density in 47 patients [14]
Figure 3: Level of each blood marker in active disease relative to controls, from the same cohort [14]
Figure 3: Level of each blood marker in active disease relative to controls, from the same cohort [14]

Reported performance and

Compartment What it measures

limitation

Oesophageal Peak eosinophil count The reference standard, but patchy biopsy, histology per high-power field and observer-dependent [8,9,10] Oesophageal Expression of 96 96% sensitivity, 98% specificity; biopsy, transcript disease-associated works on archival tissue [13] panel genes Correlates at r = 0.32 to 0.56; Eosinophil count and Peripheral blood monitoring rather than diagnosis granule proteins [14] Table 2: What each compartment measures and how well it performs The reference standard is part of the problem: Every accuracy figure above is measured against the eosinophil count, which is itself unreliable. Agreement between pathologists assessing the same biopsies is imperfect [10], and a validated scoring system incorporating basal zone hyperplasia, eosinophil abscesses, surface layering and other features was shown to outperform the peak count for both diagnosis and monitoring [9,11].

This creates an interpretive difficulty that is rarely stated. A molecular panel reported at 96 per cent sensitivity against an imperfect standard may be performing better than that figure suggests, since some of its apparent errors will be cases where the count was wrong. The 34 histologically ambiguous samples included in the validation cohort are precisely that group [13].

Correlates

Marker Status with tissue

Absolute eosinophil Best blood marker; monitoring only r = 0.56 count [14] Eosinophil-derived r = 0.54 Comparable to eosinophil count [14] neurotoxin Weak in blood, dominant in tissue Eotaxin-3 r = 0.32 [12,14] Eotaxin-1 and No correlation Not raised in active disease [14] eotaxin-2 Interleukin 5 No correlation Not raised in active disease [14] Faecal neurotoxin No correlation Not raised in active disease [14]

Table 3: Candidate blood and stool markers and their reported relationship to tissue eosinophilia [14]
Table 3: Candidate blood and stool markers and their reported relationship to tissue eosinophilia [14]

The striking result in this literature is that a transcript panel measured on the same biopsy outperforms the eosinophil count that defines the disease, and does so on archival material with no loss of accuracy [13]. That combination, better than the reference standard and compatible with tissue already in the block, is unusual. It also detects inflammation remote from the sampled site, which addresses the single most awkward feature of the histological diagnosis [8].

Blood is a different story and the title of this field invites the wrong inference. Absolute eosinophil count and eosinophil-derived neurotoxin do rise in active disease and do correlate with tissue eosinophilia, but at correlations near 0.55 they explain well under half the variance [14]. That is a monitoring signal in a patient whose diagnosis is settled, not a replacement for endoscopy in one whose diagnosis is not.

Why eotaxin-3 behaves differently in the two compartments

The most instructive detail is that CCL26 is the most highly induced transcript in oesophageal tissue [12] yet the weakest of the three correlating blood markers, at r = 0.32 [14]. A chemokine that acts locally to recruit eosinophils into the mucosa need not appear proportionally in the circulation, and the discrepancy is a reminder that a marker's importance in pathogenesis does not predict its usefulness in blood.

The same logic applies in reverse to interleukin 5, which is central to eosinophil biology and yet showed no correlation with tissue eosinophilia and no elevation in active disease in the cohort reviewed [14]. Mechanistic centrality and diagnostic utility are separate properties.

Study limitations

The limitations belong to the underlying literature. The blood marker data derive from a single cross-sectional cohort of 47 paediatric patients, so the correlation estimates carry wide uncertainty and may not transfer to adults [14]. The molecular panel was developed and validated by one group, and independent replication in other laboratories is limited [8,13,21]. Every accuracy figure is anchored to an eosinophil count of imperfect reproducibility, which biases estimates in an unknown direction [9,10]. Diagnostic criteria changed materially in 2018 when proton pump inhibitor responsiveness was reclassified from a diagnostic criterion to a treatment [3,4,17], so studies published on either side of that change enrolled different populations. Treatment exposure alters both transcript and blood measurements, and cohorts differ in how much of it they permit [13,23].

Two gaps deserve naming. No study reviewed here compares the molecular panel against the validated histology score rather than against the peak count, which is the comparison that would establish whether the panel adds information beyond better-performed histology [9,13]. And no prospective study has tested whether monitoring by blood markers rather than by repeat endoscopy changes any patient outcome [14,22].

Conclusion

Eosinophilic oesophagitis has a conserved oesophageal transcript signature spanning about one per cent of the genome, with eotaxin-3 as its most induced component. A 96-gene panel drawn from that signature identifies the disease with roughly 96 per cent sensitivity and 98 per cent specificity, retains that accuracy on archival paraffin material, separates active disease from remission and from reflux, and detects inflammation the sampled biopsy missed. Peripheral blood markers rise in active disease but correlate with tissue eosinophilia only moderately, and eotaxin-3, the dominant transcript in tissue, is the weakest of them in blood. The tissue and the blood are answering different questions, and the evidence supports using the first for diagnosis and the second, at most, for follow-up.

Acknowledgements

The transcriptome profiling cohorts and the consensus documents drawn on throughout were the work of other groups, and the credit for them is not ours.

Funding

This piece of work carried no grant, no sponsorship and no support from any institution.

Conflict of Interest

None of the authors stands to gain from anything reported here.

Ethics Statement

What is reviewed here has all appeared in print before now. No patient was approached and no fresh data collected, placing the work outside the scope of ethical review.

References

  1. Navarro P, Arias Á, Arias-González L, Laserna-Mendieta EJ, Ruiz-Ponce M, Lucendo AJ. "Systematic review with meta-analysis: the growing incidence and prevalence of eosinophilic oesophagitis in children and adults in population-based studies". Aliment Pharmacol Ther 49(2019):1116-1125.
  2. Dellon ES, Jensen ET, Martin CF, Shaheen NJ, Kappelman MD. "Prevalence of eosinophilic esophagitis in the United States". Clin Gastroenterol Hepatol 12(2014):589-596.
  3. Dellon ES, Liacouras CA, Molina-Infante J, Furuta GT, Spergel JM, Zevit N, et al. "Updated international consensus diagnostic criteria for eosinophilic esophagitis: proceedings of the AGREE conference". Gastroenterology 155(2018):1022-1033.
  4. Spergel JM, Dellon ES, Liacouras CA, Hirano I, Molina-Infante J, Bredenoord AJ, Furuta GT. "Summary of the updated international consensus diagnostic criteria for eosinophilic esophagitis: AGREE conference". Ann Allergy Asthma Immunol 121(2018):281-284.
  5. Furuta GT, Liacouras CA, Collins MH, Gupta SK, Justinich C, Putnam PE, et al. "Eosinophilic esophagitis in children and adults: a systematic review and consensus recommendations for diagnosis and treatment". Gastroenterology 133(2007):1342-1363.
  6. Liacouras CA, Furuta GT, Hirano I, Atkins D, Attwood SE, Bonis PA, et al. "Eosinophilic esophagitis: updated consensus recommendations for children and adults". J Allergy Clin Immunol 128(2011):3-20.
  7. Lucendo AJ, Molina-Infante J, Arias Á, von Arnim U, Bredenoord AJ, Bussmann C, et al. "Guidelines on eosinophilic esophagitis: evidence-based statements and recommendations for diagnosis and management in children and adults". United European Gastroenterol J 5(2017):335-358.
  8. Rothenberg ME, et al. "Diagnostic merits of the eosinophilic esophagitis diagnostic panel from a single esophageal biopsy". J Allergy Clin Immunol 149(2022):782-787.
  9. Collins MH, Martin LJ, Alexander ES, Boyd JT, Sheridan R, He H, et al. "Newly developed and validated eosinophilic esophagitis histology scoring system and evidence that it outperforms peak eosinophil count for disease diagnosis and monitoring". Dis Esophagus 30(2017):1-8.
  10. Warners MJ, Ambarus CA, Bredenoord AJ, Verheij J, Lauwers GY, Walsh JC, et al. "Reliability of histologic assessment in patients with eosinophilic oesophagitis". Aliment Pharmacol Ther 47(2018):940-950.
  11. Collins MH, Dellon ES, Katzka DA, Hirano I, Williams J, Lan L. "Eosinophilic esophagitis histology remission score: significant relations to measures of disease activity and symptoms". J Pediatr Gastroenterol Nutr 70(2020):598-603.
  12. Blanchard C, Wang N, Stringer KF, Mishra A, Fulkerson PC, Abonia JP, Jameson SC, Kirby C, Konikoff MR, Collins MH, Cohen MB, Akers R, Hogan SP, Assa’ad AH, Putnam PE, Aronow BJ, Rothenberg ME. "Eotaxin-3 and a uniquely conserved gene-expression profile in eosinophilic esophagitis". J Clin Invest 116(2006):536-547.
  13. Wen T, Stucke EM, Grotjan TM, Kemme KA, Abonia JP, Putnam PE, Franciosi JP, Garza JM, Kaul A, King EC, Collins MH, Kushner JP, Rothenberg ME. "Molecular diagnosis of eosinophilic esophagitis by gene expression profiling". Gastroenterology 145(2013):1289-1299.
  14. Konikoff MR, Blanchard C, Kirby C, Buckmeier BK, Cohen MB, Heubi JE, Putnam PE, Rothenberg ME. "Potential of blood eosinophils, eosinophil-derived neurotoxin, and eotaxin-3 as biomarkers of eosinophilic esophagitis". Clin Gastroenterol Hepatol 4(2006):1328-1336.
  15. Davis BP, Rothenberg ME. "Mechanisms of disease of eosinophilic esophagitis". Annu Rev Pathol 11(2016):365-393.
  16. Eluri S, Dellon ES. "Proton pump inhibitor-responsive oesophageal eosinophilia and eosinophilic oesophagitis: more similarities than differences". Curr Opin Gastroenterol 31(2015):309-315.
  17. Lucendo AJ, Arias Á, Molina-Infante J. "Efficacy of proton pump inhibitor drugs for inducing clinical and histologic remission in patients with symptomatic esophageal eosinophilia: a systematic review and meta-analysis". Clin Gastroenterol Hepatol 14(2016):13-22.
  18. Straumann A, Bauer M, Fischer B, Blaser K, Simon HU. "Idiopathic eosinophilic esophagitis is associated with a T(H)2-type allergic inflammatory response". J Allergy Clin Immunol 108(2001):954-961.
  19. Straumann A, Kristl J, Conus S, Vassina E, Spichtin HP, Beglinger C, Simon HU. "Cytokine expression in healthy and inflamed mucosa: probing the role of eosinophils in the digestive tract". Inflamm Bowel Dis 11(2005):720-726.
  20. Gupta SK, Fitzgerald JF, Kondratyuk T, HogenEsch H. "Cytokine expression in normal and inflamed esophageal mucosa: a study into the pathogenesis of allergic eosinophilic esophagitis". J Pediatr Gastroenterol Nutr 42(2006):22-26.
  21. Wen T, Rothenberg ME. "Clinical applications of the eosinophilic esophagitis diagnostic panel". Front Med 4(2017):108.
  22. Dellon ES. "Advances in clinical management of eosinophilic esophagitis". Gastroenterology 147(2014):1238-1254.
  23. Dellon ES, Gonsalves N, Hirano I, Furuta GT, Liacouras CA, Katzka DA. "ACG clinical guideline: evidence based approach to the diagnosis and management of esophageal eosinophilia and eosinophilic esophagitis". Am J Gastroenterol 108(2013):679-692. Citation: Nilsson H, Kowalczyk M, Lambert P. "Eosinophil-Derived Transcript Signatures in Oesophageal Biopsies and Peripheral Blood from Patients with Eosinophilic Oesophagitis: A Review." Ann Gastroenterol Dig Disord (2026) Volume 9, Issue 1 doi: 10.5281/zenodo.22207600