Abstract

Background: Burkitt lymphoma (BL) is a highly aggressive lymphoma that can be cured in up to 80% of patients when treated with intensive multi-agent chemotherapy. The distinction between BL and diffuse large B-cell lymphoma (DLBCL) is critical because there are important differences in their clinical management. However, the distinction can be difficult because of an overlap between DLBCL and BL in morphology, immunophenotype and cytogenetics. Previous work has shown that gene expression profiling can distinguish these entities with a high degree of certainty. Our previous work has demonstrated that microRNAs play a direct role in regulating key transcription factors in normal and malignant B cells. We investigated whether microRNA expression could be used to reliably distinguish BL from DLBCL.

Methods: Biopsy samples were collected from 104 patients with a diagnosis of either sporadic BL (N=25) or DLBCL (N=79). All cases were reviewed for pathology diagnosis and profiled for microRNA expression using microarrays. Using the 30 most highly differentially expressed microRNAs with the highest t-statistic, we applied singular value decomposition to identify the 10 most predictive microRNAs. Using those 10 microRNAs, we constructed a Bayesian predictor to distinguish BL from DLBCL. The predictor performance was tested using leave-one-out cross-validation. We further applied gene expression profiling to 52 cases of DLBCL to identify the molecular subsets of DLBCL: activated B cell type and germinal center B cell type DLBCL. We constructed a Bayesian predictor to distinguish these molecular subsets based upon their gene expression. A separate predictor was created from the microRNA profiles of these DLBCL subsets and tested using leave-one-out cross-validation. In order to understand the effects of lineage-specific microRNAs in B cell lymphomas, we applied FACS-sorting to isolate mature B cell subsets including naìˆve B cells, germinal center B cells, plasma cells and memory cells. We compared the microRNAs involved in germinal center differentiation to those expressed highly in Burkitt lymphoma.

Results: The predictor constructed based on microRNA expression was 90% accurate in distinguishing Burkitt lymphoma from DLBCL, using pathology diagnosis as the standard. The microRNA-based predictor was also over 90% accurate in the distinction of the molecular subsets of DLBCL, compared to the gold standard of gene expression-profiling. Further, we found that the Burkitt lymphoma cases consistently expressed microRNAs related to normal germinal center B cell differentiation, suggesting that they also maintain expression of B cell stage-specific microRNAs.

Conclusion: This study demonstrates that the microRNA expression profiles can be used to accurately distinguish Burkitt lymphoma from DLBCL. The ability of the predictor to identify the molecular subsets of patients with DLBCL and those with BL suggests that it will be useful in the diagnosis and management of patients with Burkitt lymphoma. Further, the patterns of microRNA expression and their target genes suggests a role for microRNAs in the pathophysiology of these tumors.

Disclosures: No relevant conflicts of interest to declare.

Author notes

Corresponding author