Activated Duodenal Mucosal Eosinophils in Children With Dyspepsia: A Pilot Transmission Electron Microscopic Study

Friesen, Craig A.*; Andre, Linda*; Garola, Robert†; Hodge, Charles*; Roberts, Charles*

Journal of Pediatric Gastroenterology & Nutrition: September 2002 - Volume 35 - Issue 3 - pp 329-333
Original Articles: Gastroenterology and Hepatology

Background: Activated eosinophils can be identified by electron microscopy (EM) Previous studies have shown EM evidence of eosinophil activation in a variety of gastrointestinal conditions associated with inflammation. The purpose of this study was to evaluate the activation state by EM of duodenal mucosal eosinophils in children who presented with dyspepsia and to determine if eosinophils are activated in patients with normal eosinophil counts on routine histology.

Methods: Twenty patients (ages 7–15 years) with dyspepsia were evaluated. All had normal gross endoscopies and Helicobacter pylori was excluded. Each patient had two endoscopic forceps biopsies taken from both the duodenal bulb (DB) and the second portion of the duodenum (DS) for routine histology to determine eosinophil counts. Two additional biopsies were taken from the adjacent mucosa of each site for EM evaluation. The eosinophil activation state was determined for each specimen and a degranulation index was calculated for DB specimens.

Results: On routine histology, peak eosinophil counts were greater than or equal to 20 per hpf in three patients, 11 to 19 per hpf in 12 patients, and less than or equal to 10 per hpf in 5 patients. All patients showed evidence of EM activation on DB specimens and 95% showed activation on DS specimens. The mean degranulation index was 50.5 + 12.0% with 65% of specimens revealing moderate (20 – 60%) degranulation and 30% of specimens revealing extensive (greater than 60%) degranulation.

Conclusions: Eosinophils present in the duodenal mucosa of children with dyspepsia are activated in a significant proportion of patients, even in those with normal eosinophil counts. The degree of degranulation is similar to that seen in other conditions where eosinophils have a pathogenic role.

Sections of *Gastroenterology and †Pathology, The Children's Mercy Hospital, Kansas City, Missouri, U.S.A.

Received January 14, 2002; accepted May 6, 2002.

Address correspondence to Craig A. Friesen, The Children's Mercy Hospital, Kansas City, Missouri, U.S.A. (e-mail:

Article Outline
Back to Top | Article Outline


Electron microscopy (EM) has been shown to provide a means of identifying activated eosinophils (1). Signs of eosinophil activation include inversion of core matrix densities with lucency of the core protein and vesiculotubular membrane formation (1). Previous studies have shown EM evidence of eosinophil activation in a variety of gastrointestinal conditions associated with inflammation (2–6).

Inflammation of the upper gastrointestinal tract may cause dyspepsia and inflammation has been implicated in the development of functional gastrointestinal disorders (7). An increase in mucosal mast cells has been reported with dyspepsia and irritable bowel syndrome (8–10). There are no previous studies evaluating the activation state of mucosal eosinophils in functional dyspepsia.

The purpose of this study was to evaluate the activation state by EM of duodenal mucosal eosinophils in children with dyspepsia and to determine if eosinophils are activated in patients with normal eosinophil counts on routine histology.

Back to Top | Article Outline



Twenty children ranging in age from 7 to 15 years (mean 10 years) were evaluated. There were 6 males and 14 females. All patients were undergoing upper endoscopy for the evaluation of dyspepsia, defined as pain or discomfort localized to the upper abdomen in accordance with published criteria (11). All patients had previously failed to respond to empiric H2 antagonist therapy. All patients had normal gross endoscopic examinations, without ulceration, erosion, and nodularity or masses, and with negative evaluations for Helicobacter pylori (histology and rapid urease testing of antral biopsies).

Back to Top | Article Outline

Light Microscopy

All patients had two grasp mucosal biopsies obtained from each of the following: lower one-third of the esophagus, gastric antrum, duodenum bulb (DB), and the second portion of the duodenum (DS). Biopsies were fixed in 10% formalin for standard processing and hematoxylin and eosin (H and E) staining. A minimum of 10 hpf were evaluated for eosinophil counts on duodenal biopsies.

Back to Top | Article Outline

Electron Microscopy

Two grasp mucosal biopsies were obtained from both the DB and DS at sites adjacent to the sites where tissue was obtained for H and E staining. Biopsies for EM were fixed in formyl glutaraldehyde, post fixed in osmium tetroxide, and stained with uranyl citrate. The tissue was then dehydrated in graded alcohols and embedded in epoxy resin. Sections were then stained with toluidine blue. These sections were cut and stained with lead citrate. Sections were examined and photographed. EMs were coded and reviewed by a single pathologist (RG) blinded to patient identity, clinical history, and routine histology. EM eosinophil activation was evaluated by previously published criteria (1). The degree of eosinophil activation was determined by calculating the degranulation index for duodenal bulb eosinophils on each patient. This is calculated as the percentage of the total granules which are activated as previously described (12). The degree of degranulation was determined as low (0–<20%), moderate (20–60%), or extensive (> 60%) as previously described (12).

The study protocol was approved by the Institutional Review Board of Children's Mercy Hospital. Informed consent was obtained from each subject's parent or guardian and assent obtained from the study subjects.

Back to Top | Article Outline


Light Microscopy

Biopsies revealed esophagitis by previously published criteria in 7 patients (13). Chronic non-specific gastritis (CG) consistent with published criteria was defined as a mononuclear infiltrate with reactive epithelial changes and/or mucin depletion (14). CG was present in 8 patients, including 4 of the patients with esophagitis. Biopsies were otherwise considered normal by usual pathologic criteria. Peak eosinophil counts for all duodenal specimens were > 20/hpf in 3 patients, 11–19/hpf in 12 patients, and ≤10/hpf in 5 patients. Mild chronic duodenitis was diagnosed in two patients on the basis of a focal increase in mononuclear cells with intraepithelial lymphocytes.

Back to Top | Article Outline

Electron Microscopy

Electron microscopy revealed numerous mucosal eosinophils. The cellular membranes of eosinophils were intact in most cells. Evidence of eosinophil activation by EM was widespread within duodenal specimens, particularly those obtained from the duodenal bulb. The percentage of duodenal biopsy sites where activated eosinophils could be identified by each criteria are shown in Table 1. The most widespread and consistent feature of activation was inversion of the relative core-matrix densities with partial or complete lucency of core protein (Fig. 1). Although vesiculotubular membrane formation (Fig. 2) and perigranular clouds of electron dense material (Fig. 2) could be identified within some eosinophil (s) in the majority of specimens, these changes were infrequently seen relative to inversion of core-matrix densities.

All eosinophils were found in proximity to intramucosal lymphocytes. In 21 of 22 patients, eosinophils were seen in direct apposition to a lymphocyte.

A mean number of 3.7 duodenal bulb eosinophils and 60.3 granules per patient were evaluated to determine the degranulation index. The mean degranulation index was 50.5 + 12.0% (range 5–67%). All but one patient had a degranulation index greater than or equal to 27%. The percentage of patients in each category of degranulation level are shown in Table 2. The duodenal bulb degranulation index was 53.9% in patients who had normal biopsies and 50.3% in patients with chronic non-specific gastritis (P > .05).

Back to Top | Article Outline


Childhood functional dyspepsia (FD) has been defined by Rome II as persistent or recurrent pain or discomfort in the upper abdomen with no evidence that organic disease is likely to explain symptoms (11). In the adult population, functional dyspepsia has included patients without ulcers who have histologic gastritis or duodenitis and even those with H. pylori-positive gastritis. The inclusion of these patients is felt appropriate because of the uncertain clinical relevance of these histologic findings. Healthy adults have such histologic inflammation, and, in the case of H.pylori, eradication of infection may not result in symptomatic improvement in adults with dyspepsia (15–17). Clinical correlations (e.g., response to therapy) are lacking for many common forms of inflammation. Patients with normal gross endoscopy with histologic chronic gastritis were included in this study as they are considered part of FD in adults and it has unproven clinical relevance in children. Previously we were unable to demonstrate any effect of chronic gastritis on gastric emptying or in inducing gastric dysrhythmia (18). We also included patients with normal gross esophageal endoscopy with histologic esophagitis because in all but one patient, this diagnosis was made solely on the basis of intraepithelial eosinophils. A current challenge with regard to the study of dyspepsia is to determine what (if any) types of inflammatory cells or forms of inflammation are playing a role in the generation of symptoms. The intent of the current study is to perform an initial evaluation of duodenal mucosal eosinophils with regard to activation state to help indicate whether eosinophils are potentially contributing to the process.

Electron microscopy and immunohistochemistry have provided evidence of eosinophil activation in a variety of conditions affecting the gastrointestinal tract, including esophagitis, eosinophilic gastroenteritis, Celiac disease, Crohn's disease, ulcerative colitis, and collagenous colitis (2–6,19,20). There are no previous EM studies of duodenal eosinophils in functional dyspepsia in children. By EM criteria, we found numerous activated eosinophils in all specimens (20/20) from the duodenal bulb and in 95% (18/19) of specimens from the second portion of the duodenum. The main indicator of activation was inversion of the relative core-matrix densities of the biocompartmental specific eosinophil granules with partial or complete lucency of the crystalline cores. Eosinophils contained a complement of empty and partially empty specific granules in what has been termed piecemeal degranulation (1). Additional findings consistent with activation were seen in the majority of specimens and included vesiculotubular formation and the presence of perigranular clouds of electron dense material. The degree of eosinophil degranulation in duodenal bulb mucosal biopsies of the patients in this study was moderate in 65% and extensive in 30%. The level of degranulation was more than five-fold greater than that previously reported for inflammatory bowel disease and was similar to that seen in other diseases where eosinophils are believed to have a pathogenic role (12). This would suggest that eosinophil activation may play a role in the development of dyspepsia in children. Ultimately, the clinical relevance of mucosal eosinophil activation in children will likely only be determined in randomized clinical trials of therapies directed at eosinophils.

Although eosinophil degranulation by immunohistochemistry appears to occur frequently in small intestinal mucosa from normal adults, eosinophil activation by EM criteria does not appear to occur in normal mucosa. Kato et al. (21) report that eosinophil degranulation as evidenced by extracellular major basic protein (MBP) deposition by immunohistochemistry was a remarkable finding in normal intestinal mucosa, though significantly less prevalent than in diseased tissue. This finding was more prominent in specimens obtained by forcep biopsy presumably due to tissue trauma with eosinophil cytolysis. By contrast, the findings in this and other EM studies have been a decreased core density with intact eosinophil membranes. Activation was previously demonstrated in involved tissue in a patient with eosinophilic gastroenteritis which contrasted with eosinophils from macroscopically normal tissue in this patient (3). Desremauk et al. (4) demonstrated degranulation by EM in patients with Celiac disease, but found normal granule structure in jejunal biopsies from 8 control patients with untreated irritable bowel syndrome.

There are no universally accepted criteria for the diagnosis of eosinophilic gastroenteritis in children. Lowichik and Weinberg (22) reported frequent eosinophilia on duodenal specimens from autopsies with counts > 20/hpf in 18% of specimens. It is somewhat difficult to interpret those findings as patients could not be interviewed for current gastrointestinal symptoms or followed to see if biopsy findings predisposed to symptoms. Kokkonen et al. (23) has defined mucosal eosinophilia of 6–10 eosinophils/hpf defined as mildly increased and >10/hpf as highly increased. Whitington and Whitington reported <10 eosinophils/hpf in all of 180 mucosal biopsies from control patients (24). In an attempt to determine whether eosinophils could be activated in specimens with normal eosinophil counts, we elected to take a conservative approach by reporting peak eosinophil counts. This approach was also taken in part due to the lack of universally accepted criteria and the observation that eosinophil counts can vary significantly from one hpf to the next within the same biopsy specimen. We found evidence of eosinophil activation even in patients with peak eosinophil counts of less than 10/hpf. This suggests that eosinophil counts obtained by routine histology may not give an adequate indication of potential eosinophil involvement in disease processes. Indeed, Erjefalt et al. (12) found that the level of degranulation varies significantly between diseases where the eosinophil density on routine histology does not.

Eosinophils may be activated by a variety of cytokines and inflammatory mediators which include (but are not limited to) platelet activating factor, anaphylatoxins (e.g., C5a), RANTES, MCP-2, MCP-3, MCP-4, and eotaxin (25). Eotaxin may be of particular importance in gastrointestinal eosinophilia as it has been found in mice to be important in both normal homing of eosinophils to the gastrointestinal tract and eosinophil recruitment in eosinophilic gastrointestinal inflammation. (26,27) In vitro studies in humans have shown cooperation between eotaxin and substance P (a neurotransmitter within intrinsic enteric and extrinsic afferent/nociceptive pathways) in inducing eosinophil cytotoxicity (28,29). Upon activation, eosinophils may release any of a large number of mediators with wide-ranging effects on inflammation, motility, and neural function (30). Granule proteins are generally cytotoxic and include major basic protein (MBP), eosinophil cationic protein (ECP), eosinophil peroxidase (EPO), and eosinophil-derived neurotoxin (EDN). MBP is the protein present in the granule core shown to be partially or completely lucent in the patients in the present study. ECP, EPO, and EDN have been localized to the granule matrix (31). Substance P has also been localized to intestinal eosinophils in humans (32).

Activation of eosinophils may be associated with gastric dysmotility. Abnormalities in gastric emptying and electrogastrography have been reported with gastrointestinal allergy (33,34). In a mouse model, gastrointestinal allergen exposure was associated with extensive eosinophilic inflammation and the development of gastric dysmotility (27). Mice deficient in eotaxin have reduced eosinophil infiltration and appear to be protected from the development of gastromegaly. Although not a part of this study protocol, all the patients in the current report also had a solid-phase gastric emptying scintigraphic study and EGG. An abnormal gastric emptying study (one hour emptying <30%) was present in 50% (10/20). An abnormal EGG (% normal 2–4 cpm waves <70% or pre-prandial > post-prandial power) was abnormal in 60% (12/20). Although one or the other was abnormal in 75%, the study design does not allow any cause-and-effect conclusions to be drawn, only an observation of the frequent co-existence of eosinophil activation and dysmotility.

In conclusion, duodenal eosinophil activation of a moderate to extensive degree is a very common finding in children with dyspepsia. However, future studies are needed to determine the significance of eosinophil activation in the disease process or symptom generation. EM may have value as a modality to study neuroimmune interactions in children with dyspepsia.

Back to Top | Article Outline


1. Dvorak AM, Weller PF. Ultrasound analysis of human eosinophils. Chem Immunol 2000; 76:1–28.
2. Justinich CJ, Ricci Jr., A Kalafus DA, Treem WR, Hyams JS, Kreutzer DL. Activated eosinophils in esophagitis in children: a transmission electron microscopic study. J Pediatr Gastroenterol Nutr 1997; 25:194–98.
3. Torpier G, Colombel JF, Mathieu-Chandelier C, Capron M, Dessaint JP, Cortot A, Paris JC, Capron A. Eosinophilic gastroenteritis: ultrastructural evidence for a selective release of eosinophil major basic protein. Clin Exp Immunol 1988; 74:404–8.
4. Desreumaux P, Delaporte E, Colombel JF, Capron M, Cortot A, Janin A. Similar IL-5, IL-3, and GM-CSF synthesis by eosinophils in the jejunal mucosa of patients with Celiac disease and dermatitis herpetiformis. Clin Immunol Immunopath 1998; 88:14–21.
5. Dvorak AM, Monahan RA, Osage JE, Dickersin GR. Crohn's disease: transmission electron microscopic studies. II. Immunologic Inflammatory response. Alterations of mast cells, basophils, eosinophils, eosinophils, and the microvasculature. Human Path 1980; 11:606–19.
6. Raab Y, Fredens K, Gerdin B, Hallgren R. Eosinophil activation in ulcerative colitis. Studies on mucosal release and localization of eosinophil granule constituents. Dig Dis Sci 1998; 43:1061–70.
7. Collins SM. The immunomodulation of enteric neuromuscular function: implications for motility and inflammatory disorders. Gastroenterol 1996; 111:1683–99.
8. Matter SE, Bhatia PS, Miner PB. Evaluation of antral mast cells in non-ulcer dyspepsia. Dig Dis Sci 1990; 35:1358–63.
9. Weston A, Biddle W, Bhatia P, Miner P. Terminal ileal mucosal mast cells in irritable bowel syndrome. Dig Dis Sci 1993; 38:1590–95.
10. O'Sullivan M, Clayton N, Breslin NP, Harman I, Bountra C, McLaren A, O'Morain GA. Increased mast cells in the irritable bowel syndrome. Neurogastroenterol Motil 2000; 12:449–57.
11. Rasquin-Weber A, Hyman PE, Cucchiara S, Fleisher DR, Hyams JS, Milla PJ, Staiano A. Childhood functional gastrointestinal disorders. Gut 1999; 45 (Suppl II): II60–II68.
12. Erjefalt JS, Greiff L, Andersson M, Adelroth E, Jeffery PK, Persson CGA. Degranulation patterns of eosinophil granulocytes as determinants of eosinphil driven disease. Thorax 2001; 56:341–44.
13. Hyams JS, Ricci Jr, A Leichtner AM. Clinical and laboratory correlates of esophagitis in young children. J Pediatr Gastroenterol Nutr 1988; 7:52–6.
14. Dohil R, Hassall E, Jevon G, Dimmick J. Gastritis and gastropathy of childhood. J Pediatr Gastroenterol Nutr 1999; 29:378–94.
15. Kreuning J, Bosman FT, Kuiper G, et al. Gastric and duodenal mucosa in “healthy” individuals: an endoscopic and histopathological study of 50 volunteers. J Clin Pathol 1978; 31:69–77.
16. Blum AL, Tally NJ, O'Morain C, et al. lack of effect of treating Helicobacter pylori infection in patients with nonulcer dyspepsia. N Engl J Med 1998; 339: 1875–81.
17. Talley NJ, Vakil N, Ballard ED, et al. Absence of benefit of eradicating Helicobacter pylori in patients with nonulcer dyspepsia. N Engl J Med 1999; 341:1106–11.
18. Friesen CA, Andre L, Seidel G, et al. gastric myoelectrical activity, gastric emptying, and chronic, non-specific gastritis in children with dyspepsia. J Pediatr Gastroenterol Nutr 2001; 33:417.
19. Makiyama K, Kanzaki S, Yamasaki K, Zea-Iriarte W, Tsuji Y. Activation of eosinophils in the pathophysiology of ulcerative colitis. J Gastroenterol 1995; 30(Suppl VIII):64–9.
20. Levy AM, Yamazaki K, Van Keulen VP, Bargart LJ, Sandborn WJ, Phillips SF, Kephart GM, Gleich GJ, Leiferman KM. Increased eosinophil infiltration degranulation in colonic tissue from patients with collagenous colitis. Am J Gastroenterol 2001; 96:1522–28.
21. Kato M, Kephart GM, Talley NJ, Wagner JM, Sarr MG, Bonno M, McGovern TW, Gleich GJ. Eosinophilic infiltration and degranulation in normal human tissue. Anatom Record 1998; 252:418–25.
22. Lowichik A, Weinberg AG. A quantitative evaluation of mucosal eosinophils in the pediatric gastrointestinal tract. Mod Path 1996; 9:110–14.
23. Kokkonen J, Runska T, Karttunen TJ, Niinimaki A. Mucosal pathology of the foregut associated with food allergy and recurrent abdominal pain in children. Acta Paediatr 2001; 90:16–21.
24. Whitington PF, Whitington GL. Eosinophilic gastroenteropathy in childhood. J Ped Gastroenterol Nutr 1988; 7:379–85.
25. Elsner J, Kapp A. Activation of human eosinophils by chemokines. Chem Immunol 2000; 76:177–207.
26. Mishra A, Hogan SP, Lee JJ, Foster PS, Rothenberg ME. Fundamental signals that regulate eosinophil homing to the gastrointestinal tract. J Clin Invest 1999; 103:1719–27.
27. Hogan SP, Mishra A, Brandt EB, Royalty MP, Pope SM, Zimmermann N, Foster PS, Rothenberg ME. A pathological function for exotoxin and eosinophils in eosinophilic gastrointestinal inflammation. Nature Immunol 2001; 2:353–60.
28. El-Shazly A, Ishikawa T. Novel co-operation between eotaxin and substance P in inducing eosinophil-derived neurotoxin release. Mediators of Inflamm 1999; 8:177–79.
29. Hokfelt T, Pernow B, Wahren J. Substance P: A pioneer amongst neuropeptides. J Int Med 2001; 249:27–40.
30. Gleich GJ. Mechanisms of eosinophil-associated inflammation. J All Clin Immunol 2000; 105:651–663.
31. Peters MS, Rodriguez M, Gleich GJ. Localization of human eosinophil granule major basic protein, eosinophilic cationic protein, and eosinophil-derived neurotoxin by immunoelectron microscopy. Lab Invest 1986; 54:656–62.
32. Metwali A, Blum AM, Ferraris L, Klein JS, Fiocchi C, Weinstock JV. Eosinophils within the healthy or inflamed human intestine produce substance P and vasoactive intestinal peptide. J Neuroimmunol 1994; 52:69–78.
33. Ravelli AM, Tobanelli P, Volpi S, Ugazio AG. Vomiting and gastric motility in infants with cow's milk allergy. J Ped Gastroenterol Nutr 2001; 32:59–64.
34. Sabra A, Bellanti JA. Delayed gastric emptying, gastroesophageal reflux, and dyspeptic symptoms: The pathogenetic role of food allergy. Ped Gastroenterol Nutr 2000; 31:S206.

Cited By:

This article has been cited 3 time(s).

Journal of Pediatric Gastroenterology and Nutrition
Useful Biomarkers in Pediatric Eosinophilic Duodenitis and Their Existence: A Case-control, Single-blind, Observational Pilot Study
Neilan, NA; Dowling, PJ; Taylor, DL; Ryan, P; Schurman, JV; Friesen, CA
Journal of Pediatric Gastroenterology and Nutrition, 50(4): 377-384.
PDF (632) | CrossRef
Current Opinion in Gastroenterology
Pediatric gastric and duodenal disorders
Blank, C; Sabri, M; Di Lorenzo, C
Current Opinion in Gastroenterology, 20(6): 551-555.

PDF (73)
Journal of Pediatric Gastroenterology and Nutrition
Clinical Efficacy and Pharmacokinetics of Montelukast in Dyspeptic Children with Duodenal Eosinophilia
Friesen, CA; Kearns, GL; Andre, L; Neustrom, M; Roberts, CC; Abdel-Rahman, SM
Journal of Pediatric Gastroenterology and Nutrition, 38(3): 343-351.

PDF (469)
Back to Top | Article Outline

Dyspepsia; Eosinophilia; Electron microscopy

© 2002 Lippincott Williams & Wilkins, Inc.