The present study of HIS revealed: 1) a relatively high sEo count, 2) the presence of iNeu and iEo, and 3) the presence of LFs. In part, the above findings paralleled the results of a Swedish report in which biopsy samples from HIS cases revealed the presence of both mucosal eosinophilia and lymphoid follicles . In that Swedish study, the authors showed mild mucosal eosinophilia (mean 30 cells/mm2) in 17 HIS cases and also suggested an association of HIS with irritable bowel syndrome . However, HIS may be present in specimens containing polypoid or neoplastic lesions, such as conventional adenomas or hyperplastic polyps , and the 17 HIS cases in the Swedish study included 3 cases having tubular adenomas, 3 hyperplastic polyps, and 2 diverticular disease. These accompanying lesions may be considered to modify the cell types and numbers present. Thus, in the present study, we excluded specimens containing polypoid or neoplastic lesions. Nevertheless, even using these selected specimens, we could confirm the results of the Swedish study .
First, the average sEo value in the present specimens was slightly over 20 cells/HPF, which possibly meets the histologic criteria for eosinophilic gastroenteritis (EGE) , although the eosinophil number in normal colonic mucosae remains in discussion . EGE is characterized by an accumulation of eosinophils in the digestive tract, and sometimes requires histologic confirmation. The pathogenesis of EGE is thought to be related to a hypersensitivity reaction, and by radiology and endoscopy, mucosal edema may be seen in EGE . Interestingly, marked mucosal edema has been reported in HIS, too . Concerning EGE criteria, accompanying infectious disease should be excluded from its diagnosis . However, from the results of the present study, some HIS cases may possibly be diagnosed incorrectly to be EGE because a histologic sign of HIS (viz. the fringe formation on the mucosal surface) may be too subtle or focal for it to be easily recognized, even by experienced pathologists. Moreover, the diagnostic fringes are not always found everywhere in the large intestine of HIS cases  and we noted that stromal or intraepithelial eosinophil infiltration (sEo or iEo, respectively) was not influenced by the degree of fringe distribution. In the present study, intraepithelial eosinophil counts were related not only to their stromal counts, but also to both the intraepithelial and stromal neutrophil counts. Thus, stromal and intraepithelial eosinophilic infiltration might be a central indicator of HIS and a host reaction to HIS.
HIS might be underestimated, indeed considered a harmless condition that is not worthy of being called a “disease”, since spirochetal residence may be part of the normal flora in human large intestines [1, 3]. However, neutrophilic infiltration, especially in the surface epithelial layer, generally indicates active mucosal inflammation, and its presence in HIS indicates that HIS exhibits histologic signs of “infectious disease”. In particular, iEo was observed in more than half the present specimens. In the present study, we found that neutrophilic and eosinophilic infiltration into the surface epithelial layer (viz. iNeu and iEo, respectively) were strongly correlated, while our data suggesting that intraepithelial neutrophil infiltration, at least, might be a reaction to spirochetal attachments on the surface epithelium and/or spirochetal presence in crypt lumens. In other words, these spirochetal burdens might be thought sufficient stimuli to induce intraepithelial neutrophilic or eosinophilic infiltration. However, these active inflammation parameters were not influenced by whether intraepithelial spirochetal invasion was or was not present. One possibility to consider is that spirochetal invasion into the epithelial layer might not be destructive or cytotoxic to the surface epithelium, but simply permeative (mainly between the cell borders of intact or apoptotic surface epithelium). If that is correct, spirochetal invasion might not add further stimuli to those due to the spirochetal burden on the epithelial surface and/or crypt lumens. Similarly, the degree of eosinophilic activation and tissue damage were presumably not severe enough to increase mast cells since the mast cell count in the present study was not related to sEo values, and moreover remained lower or similar to the basal level of mast cells shown in previous reports [14, 15].
In the present study, LFs were observed in half of all specimens from HIS cases. These findings suggest the presence of HIS-related chronic mucosal inflammation, although the presence of LFs did not influence eosinophils, neutrophils, or mast cells counts or the spirochetal infection parameters that we examined. In our previous immunohistochemical investigation using surgically removed specimens with HIS, possible spirochetal fragments were detected within LFs . On that basis, LFs might be another immunologic host reaction to a spirochetal burden.
The present study has some limitations regarding its materials: (1) we analyzed a small number of HIS cases that were archived and that were detected at only a single hospital (although it is a medical center in a location neighboring Tokyo, the capital of Japan, and specimens were also collected from other hospitals in the same prefecture) in only three calendar years, (2) potential inaccuracy may have resulted from IHC cross-reaction with the antibody used in the present study (antibodies specific for Brachyspira species not being commercially available) and we did not perform further confirmation, such as species genotyping or in situ hybridization (cases were limited to those exhibiting distinctive fringes), and (3) the inflammation parameters were limited to a few cell types, and excluded lymphocytes, plasma cells, and macrophages (these cells, especially macrophages, might contribute to the mucosal immunity to tissue-invading spirochetes , although examining them was beyond the scope of present study).