Search

Inflammatory fibroid polyps (IFPs) are rare mesenchymal tumors of the gastrointestinal tract that consist of spindle-shaped stromal cells and an inflammatory infiltrate rich in eosinophils. Their etiology and histogenesis remain unknown. Based on previous reports of their immunoreactivity for CD34 and c-kit biomarkers, IFPs have been thought to be related to gastrointestinal stromal tumors (GISTs). After reviewing the current literature and examining IFPs at the light microscopic level, we evaluated a series of IFPs using an extensive panel of immunohistochemical and in situ hybridization markers in an effort to gain insight into their etiology and histogenesis and to determine their true relationship to GISTs. Sixteen routinely processed IFP specimens (14 gastric, 1 ileal, and 1 rectal) were immunohistochemically stained for antibodies to CD34, HMB-45, desmin, smooth muscle actin, calponin, h-caldesmon, anaplastic lymphoma kinase, S-100 protein, epithelial membrane antigen, c-kit (CD117), stem cell factor (SCF/N19 or kit ligand), p53, bcl-2, cyclin D1, and human herpesvirus-8 (HHV8). In situ hybridization for Epstein-Barr virus-encoded RNA (EBER) was also performed. Ten cases were further evaluated for the dendritic cell markers fascin, CD21, CD23, and CD35. Stromal cells were diffusely positive for CD34 and fascin in all (100%) cases, and these stromal cells were, in addition, immunoreactive for calponin and smooth muscle actin in 88% and 25% of cases, respectively. CD35 was also found to be focally reactive in the stromal cells. Cyclin-D1 was overexpressed in all (100%) IFPs. All other immunohistochemical markers and EBER were negative in the stromal cells. These findings suggest that the proliferating stromal cells in IFPs are of dendritic cell origin, with some cases also exhibiting myofibroblastic features. Absence of c-kit, SCF, and h-caldesmon immunoreactivity fails to support a relationship to GISTs. We also conclude that Epstein Barr virus and HHV8 are unlikely etiologic agents of IFPs. Overexpression of cyclin D1 in all cases suggests that a defect in cell-cycle regulation may be involved in the growth of IFPs.

It is important to distinguish Barrett's esophagus (BE) from intestinal metaplasia related to carditis because these conditions have a different natural history, risk of malignancy, and treatment. However, the distinction between these entities is difficult both clinically and pathologically. The aim of this study was to evaluate and compare the immunostaining pattern of five mucin core polypeptides in BE to cases of carditis or antritis with intestinal metaplasia. Routinely processed mucosal biopsies from 22 patients with intestinal-type BE, 24 patients with cardia intestinal metaplasia (10 Helicobacter pylori positive), 17 patients with antral intestinal metaplasia (all H. pylori positive), 20 control patients with a normal antrum, and 22 control patients with a normal cardia were immunostained with monoclonal antibodies against MUC1, MUC2, MUC3, MUC5AC, and MUC6 mucin core polypeptides. Staining was evaluated separately for goblet cells and non-goblet columnar cells and compared between all groups. A significantly higher number of BE cases (P < 0.05) showed goblet cell staining for MUC1 (55%) or MUC6 (32%) compared with patients with carditis with intestinal metaplasia (MUC1 14%, MUC6 7%) or antritis with intestinal metaplasia (MUC1 6%, MUC6 0%). BE also showed a higher frequency of MUC1 and MUC6 positivity in non-goblet columnar cells compared with carditis and antritis cases with intestinal metaplasia. Only cases of BE showed combined MUC1 and MUC6 staining (sensitivity 23%, specificity 100%). The sensitivity and specificity of MUC1 staining for BE are 55% and 96%, respectively, and for MUC6 staining 30% and 96%, respectively. Interestingly, normal gastric cardia mucosa also showed a significantly higher prevalence of MUC2 and MUC3 expression in glandular epithelium (29% and 38%, respectively) compared with the antrum (0% for both markers) (P < 0.05). In conclusion, MUC1 and MUC6 expression in BE is distinct from that of the cardia and antrum with intestinal metaplasia; thus, immunophenotyping for these markers may have some value in a subset of patients in helping to separate BE from patients with intestinal metaplasia of the cardia. Columnar epithelium in the "normal" gastric cardia has a partially intestinalized phenotype and, as a result, may represent an early form of metaplastic epithelium.

Angioimmunoblastic lymphoma (AIL) is a T-cell proliferation with a distinct clinical presentation that often poses a difficult diagnostic challenge. Angioimmunoblastic lymphoma is characterized by prominent vascular and stromal proliferations. Using a panel of antibodies, we investigated the nature of the stromal component in 15 cases of AIL as compared with 40 cases of nodal-based peripheral T-cell lymphoma (PTCL) of other types. As has been previously noted, extrafollicular proliferations of follicular dendritic cells (detected by CD21 and low molecular-weight nerve growth factor receptor staining) were highly associated with AIL and were only rarely seen in other lesions. Unexpectedly, large networks of desmin-positive reticulum cells also were noted in all cases of AIL evaluated. These cells with characteristic long cytoplasmic processes were present in much smaller numbers or only rarely in other types of peripheral T-cell lymphoma. This population of nodal stromal cells, a subset of the fibroblastic reticulum cells detected by vimentin immunostaining, may be responsible for the prominent reticulum deposition seen in AIL. No association of AIL with proliferations of other types of reticulum cells (e.g., interdigitating dendritic cells or histiocytes) was noted. These findings suggest that networks of follicular dendritic and desmin-positive reticulum cells are useful diagnostic features in angioimmunoblastic lymphoma and probably are related to the pathogenesis of this entity.