Molecular and immunohistochemical characterization of B-cell lymphoma-2-negative follicular lymphomas

Aberrant expression of the antiapoptotic protein BCL (B-cell lymphoma)-2 in neoplastic germinal centers is one of the diagnostic hallmarks of follicular lymphoma. If BCL-2 cannot be detected by immunohistochemistry, the distinction between florid follicular hyperplasia and follicular lymphoma might become a diagnostic challenge. Most of those cases also lack the typical t(14;18), and the underlying pathophysiologic conditions of follicular lymphoma that lack BCL-2 protein expression are largely unknown. Here, we collected 18 BCL-2-negative follicular lymphoma cases from 5 different institutions. After restaining, 9 cases proved to be truly BCL-2 negative (6 follicular lymphoma grade 2, 2 follicular lymphoma grade 3a, and 1 follicular lymphoma grade 3b). In 4 additional cases, BCL-2 was very faint (all grade 2). Of the 9 BCL-2-negative follicular lymphoma cases, 2 were negative for CD10 (22%); all showed expression of BCL-6. Apoptotic level as determined by caspase 3 was the lowest in the BCL-2-positive follicular lymphoma group (15 ± 8 mm(2)), the highest in the normal/reactive group (n = 7, 60 ± 12 mm(2)) and very similar in the BCL-2 low follicular lymphoma and BCL-2-negative follicular lymphoma groups (25 ± 13 and 33 ± 19 mm(2), respectively), assuming an intermediate position between reactive follicles and BCL-2-positive neoplastic follicles (P < .001 [Kruskal-Wallis]). Also noted was a difference in proliferation fractions between the BCL-2-positive follicular lymphoma (27% ± 15%), the BCL-2 low follicular lymphoma (30% ± 20%) and the BCL-2-negative follicular lymphoma groups (30% ± 22%). Regarding the network of follicular dendritic cells, 8 (89%) of 9 cases from the BCL-2-negative subgroup showed disrupted, weakly developed networks, whereas all of the follicular lymphoma BCL-2 low-expression cases showed a well-defined and strongly developed follicular dendritic cell network. Among BCL-2-negative follicular lymphoma, BCL-2 and BCL-6 breaks were found in 1 case each, whereas in the follicular lymphoma BCL-2 low group, only 1 case with a BCL-6 break was recorded. No statistically significant result was achieved upon assessment of BCL-2α or BCL-2β RNA or the ratio of α/β isolated by real-time-polymerase chain reaction. Taken together, BCL-2-negative follicular lymphoma did not show a BCL-2 break on the genetic level and showed both increased apoptotic and proliferation rates compared with BCL-2-positive follicular lymphoma. In our series, BCL-6 breaks were infrequent in BCL-2-negative follicular lymphoma.