Four cases of gastric carcinoma are described that are associated with an osteoclast-like giant cell (OGC) stromal component. The patients were all middle-aged men (range 53-63 years). Microscopically, the tumors were characterized by a bland cytologic appearance, and an either solid or cribriform pattern. Osteoclast-like giant cells were found adjacent to, or intimately intermixed with, the neoplastic cells in the primary gastric masses and in the lymph nodal metastases and were often associated with lymphocytes, histiocytes, and desmoplastic stroma. By immunohistochemistry, mononuclear cells and OGCs showed diffuse positivity for alpha-1-antichymotrypsin, alpha-1-antitrypsin, and CD68. Neoplastic cells that were positive for keratin and CEA, also showed reactivity for vimentin and the latent membrane protein of Epstein-Barr virus in one case. At follow-up, three patients had died at 13, 15, and 24 months after diagnosis, and one is still alive, without evidence of disease, after 120 months. This report describes a novel variant of gastric carcinoma with distinctive and histologic features.

We analyzed individual VH gene rearrangements in 55 consecutive B-chronic lymphocytic leukemia (B-CLL) patients collected from a northeastern region of Italy, stressing the possible differences related to geographic characteristics of the cohorts studied. Considering the percentage of somatic mutations present in the VH gene sequences and using the 98% cut-off value, 38 of the 55 B-CLL (69%) patients displayed somatic hypermutations and 17 (31%) had a germline configuration. Our results confirm and extend the observations of a bias in the use of certain VH, DH, and JH genes among B-CLL cells. The most frequently used VH genes were VH1-69 (12.7%) with VH3-23 (12.7%) and VH4-34 (10.9%). Collectively these genes accounted for 36.3% of the cases. In the mutated cases, the range of mutations varied from 2% to 15%, with a median of 6.5%. VH1-69 (7 cases, all unmutated) carried few mutations as opposed to VH3-23 (7 cases, 5 of which mutated), VH4-34 (6 cases, all mutated), and VH3-30 (5 cases, all mutated), which show a high load of mutations. D3 family genes were found frequently (38.1%) followed by D2 (27.2%) and D6 (18.1%). The individual D segment most frequently used was D3-3, which was present in 16.3% of cases. There was predominance of the JH4 gene (49%) followed by JH6 (40%). Analysis of the distribution of replacement and silent mutations in the mutated sequences using the method of Lossos showed in 39.4% of cases evidence of antigen selection in the framework region and/or complementary determining regions. In comparison with a recent study on B-CLL patients from the Mediterranean area, the VH4-34 gene was significantly overused in the mutated group at a percentage double that of the Italian cohort reported in this study (10.9% vs. 5%), but at a frequency similar to the entire Mediterranean region (10.7%). We also found an over-representation of VH1-69 usage in the germline group, at a frequency (12.7%) higher than previously described by the same authors (Italian 8%, Mediterranean 10.7%). On the contrary, VH3-07 and VH3-49 were not much used in our study (5.4% and 1.8%, respectively) compared with the Italian group (8% and 5.1%). In our study, VH3-23 gene segment was frequently expressed, at frequency as high as that of VH1-69, a finding in keeping with reported B-CLL Italian data, but higher than the entire series of the Mediterranean area (12.7% vs. 9.2%); VH3-21 gene, frequently expressed in northern European CLL but rarely in the Mediterranean area, was completely absent. This biased usage of VH family genes may reflect a geographic leukemic repertoire, perhaps owing to a peculiar genetic background, depending on variations in germline composition of the IgVH locus or to the effect of a potential environmental element less frequently encountered in different regions.

No abstract