Comparing MicroRNA Expression In Aggressive and Indolent Non-Hodgkin Lymphomas Identifies a Prognostic Signature for Mantle Cell Lymphoma

Abstract 800

Mantle cell lymphoma (MCL) is a B-cell non-Hodgkin lymphoma (NHL) accounting for ~6% of all NHL. It is sensitive to combination chemotherapy, but remission durations are short without approaches such as stem cell transplantation (SCT). Most patients are incurable, but the clinical course is variable, with some patients succumbing quickly, while others survive >10 years. MicroRNAs (miRs) are small, non-coding RNAs that regulate gene expression by inhibiting mRNA translation. miRs are useful in the prognostic assessment of tumors, but work to date examining differences between MCL and normal lymphoid tissues, have only identified 2 miRs involved in MCL prognosis (Zhao JJ, Blood, 2010; Di Lisio L, Leukemia, 2010). We used a novel approach to identify a prognostic miR signature in MCL. We hypothesized that a miR signature defining aggressiveness can be obtained by comparing miR expression profiles of aggressive NHL with indolent NHL, and that this signature when applied to a set of MCL cases, may aid in MCL prognosis. Total RNA was extracted from 135 formalin-fixed paraffin-embedded samples obtained at primary diagnosis (Table 1). RNA from a training set of 19 indolent and 20 aggressive NHL cases was analyzed on a high-throughput quantitative real-time PCR (qRT-PCR) platform assessing the expression of 365 miRs and 3 endogenous controls (TaqMan Human MicroRNA Array v1.0: TLDA, ABI) using the DDCt method. A two-sample Wilcoxon Rank sum test corrected for false discovery rate was used to assess the significance of differential expression for each miR between aggressive and indolent NHL. The 14 most significantly differentially expressed miRs (p<0.001, FDR<0.02) were validated on an independent set of 25 indolent NHL and 19 aggressive NHL by qRT-PCR, and analyzed using the DDCt method. Univariate analysis using a one-sided t-test yielded 9 miRs that validated on the independent NHL set. Multivariable analysis demonstrated the ability of this 9 miR signature to distinguish between aggressive and indolent NHL (p<0.0001). Applying this signature to a set of 32 MCL patients with complete outcome data (Table 2) separated a poor prognosis group (median OS: 15 months, range: 4–40 months) from a good prognosis group (median OS: 88 months, range: 41–131 months) (Fig. 1). Among the 9 miRs were miR-29c, shown to have some prognostic value in MCL by Zhao et al., and miR-26a, shown to be important in MCL pathogenesis by Di Lisio et al. In light of the overlap with such recent studies, we believe the 9 miR prognostic signature we have identified may be of clinical utility. We are currently identifying mRNA targets for this miR signature and validating both the signature and the deregulated expression of these targets on a larger set of 200 MCL samples with known outcome data.

Fig. 1.

View largeDownload slide

Psrincipal component analysis demonstrating separation of MCL cases into a good prognosis group in red (median OS: 88 months, range: 41–131 months) and a poor prognosis group in blue (median OS: 15 months, range: 4–40 months) based on expression of a 9 miR aggressiveness signature.

Fig. 1.

View largeDownload slide

Psrincipal component analysis demonstrating separation of MCL cases into a good prognosis group in red (median OS: 88 months, range: 41–131 months) and a poor prognosis group in blue (median OS: 15 months, range: 4–40 months) based on expression of a 9 miR aggressiveness signature.

Close modal