Targeting of TNFAIP3 (A20) Links MiR-125a and MiR-125b to Constitutive Activation of the NF-KB Pathway and Lymphomagenesis

Abstract 774

The most aggressive subtype of diffuse large B-cell lymphoma (DLBCL), ABC-DLBCL, is characterized by constitutive activation of NF-κB signals, in association with somatic mutations in regulators of this pathway. However, in a large fraction of cases the molecular basis for the elevated NF-κB activity is unclear. MicroRNAs (miRNA) play a central role in immune function, in part by modulating the NF-κB system; yet, their role in the NF-κB deregulation found in DLBCL is unknown.

To address this issue, we created a copy number/expression map of miRNAs in ∼100 DLBCLs, and integrated these data with miRNA target prediction models. This supervised strategy showed that miR-125a and miR-125b, which we found to be amplified/overexpressed in ∼ 30% of DLBCLs, are predicted to target TNFAIP3 (A20) via two fully conserved binding sites. A20 is an ubiquitin editing enzyme that negatively regulates NF-κB complexes and acts as a lymphoma suppressor gene, suggesting that its abnormal downregulation by miRs-125a/b may impinge on DLBCL pathogenesis.

To validate these initial findings, we stably expressed miR-125a and miR-125b in multiple DLBCL cell lines, and used western blots to show that both miRNAs downregulate A20. In parallel assays, we used anti-miR oligos and miRNA sponge constructs in cell lines expressing high levels of miRs-125a/b, and found that their inhibition upregulates A20 levels. To study the interaction between miRs-125a/b and A20, we created reporter constructs with A20 sequences wild-type (WT) or mutated for the miR-125 binding sites; luciferase activity of the WT (but not the mutant) constructs was inhibited by miRs-125a/b confirming that they directly target A20.

We hypothesized that miRs-125a/b-mediated inhibition of A20 activates NF-κB in DLBCL. Indeed, phosphorylation and degradation of the NF-κB inhibitory protein IκBα was markedly enhanced in DLBCL cell lines ectopically expressing miRs-125a/b, in comparison to their isogenic counterparts lacking these miRNAs. Further, nuclear accumulation of RelA was also elevated in miRs-125a/b expressing cells. Finally, we used a NF-κB reporter construct, to show that DLBCL cells stably expressing miRs-125a/b had significantly higher NF-κB activity (p<.005) than those lacking these miRNAs. In key additional assays, we found that stable expression of miRs-125a/b-specific sponges blocked the phosphorylation/degradation of IκBα in association with A20 upregulation.

Each miRNA targets multiple genes. Thus, to confirm that the effects of miRs-125a/b on NF-κB activity result mainly from A20 inhibition, we stably expressed an A20 construct lacking miRs-125a/b binding sites in DLBCL cell lines genetically modified to overexpress these miRNAs. The ratios of phosphorylation/degradation of IκBα and the nuclear localization of RelA showed that A20 re-expression rescued the miRs-125a/b-induced activation of NF-κB in DLBCL, asserting the key role of A20 targeting on miRs-125a/b function.

MiR-125a and −125b are thought to act as oncomiRs in the hematopoietic lineage. In probing this concept, we found that DLBCLs stably expressing miRs-125a/b grow at significantly faster rate than their empty-vector counterparts (p<0.05), and have enhanced clonogenicity potential (p<0.05), tested by colony forming assays. The poor response to chemotherapy found in ABC-like DLBCL may be related to high NF-κB signals, and resistance to apoptosis. To investigate if miRs-125a/b render DLBCLs resistant to apoptosis in a therapeutic setting, we exposed cells lines expressing miRs-125a/b to H2O2, a reactive oxygen species generated by anticancer agents; we found that miRs-125a/b expressing DLBCLs were significantly more resistant to H202-induced apoptosis than isogenic cells lacking these miRNAs (20±2.8% vs. 45±3.8% apoptosis rate, p< 0.05). Finally, we examined the interplay between miRs-125a/b and A20 in primary DLBCLs, and found that tumors with the highest miRNA levels have significantly lower A20 protein expression (p<0.05, densitometric calculation of A20 levels).

Together, we defined miRNA deregulation as a novel mechanism for aberrant activation of NF-κB signals in DLBCL, and showed that A20 targeting by the lymphomagenic miRs-125a/b enhances lymphoma aggressiveness. These data bring mechanistic insight into the oncogenic activities of miRs-125a/b, and may guide the development of miRNA-based therapeutic strategies in hematological cancers.