Microrna-126 Levels Modulate B-ALL Aggressiveness but Require Tight Regulation to Guarantee Disease Maintenance

We have previously identified microRNA-126 as a central modulator of B-ALL biology (Nucera et al, Cancer Cell). Ectopic miR-126 expression in murine stem and progenitor cells is sufficient and necessary to induce and maintain a B-ALL with a Philadelphia(Ph)-like gene expression profile. Importantly, miR-126 is expressed in the majority of human Ph+ B-ALL. Interestingly, there is marked intra-tumoral heterogeneity with regards to miR-126 levels, defining a new layer of subclonal architecture different from genetic subclones. By prospectively isolating miR-126(high) and miR-126(low) subpopulations exploiting a lentiviral miRNA reporter vector, we consistently found a more aggressive behavior of the miR-126(high) subset in xenograft models. To detect miR-126-dependent pathways that may explain this more aggressive behavior, we engineered 2 primary human B-ALLs with lentiviral vector cassettes containing (1) a reverse tetracycline transactivator (rTTA) and (2) a Tet-operon driving overexpression of miR-126 (miR-126-OE), a miR-126-3p seed mutant (miR-126-OE-SM) or a miR-126 sponge (miR-126-KD) performing serial passages in mice following enrichment for transduced cells. Vector cassettes were induced in vivo by doxycycline administration to mice that have developed full blown leukemia, and blasts were collected within 72hr. In patient 1, acute miR-126-OE induced apoptosis, while it was well tolerated in the disease from patient 2 that we hence studied in more detail. RNA sequencing highlighted more than 2,000 deregulated genes between miR-126-OE and miR-126-OE-SM or miR-126-KD blasts (qValue <0.1, logFC <-0.2 or >0.2). Genes down-regulated upon miR-126-OE included transcription factors associated with B cell differentiation (PAX5, IKZ1, TCF3, BCL6), cell cycle regulators (CCND2, RBL1, E2F7, EP300) and ribosomal genes. Moreover, pathway analysis revealed significant de-regulation in JAK-STAT signaling, cell proliferation, adhesion/migration, BCR signaling, FOXO signaling and cell metabolism. Strikingly, there was a strong overlap (221 genes) with a miR-126 switch-off signature obtained from the Ph-like B-ALL mouse model. These shared genes were enriched in the categories "cell proliferation", "cancer-related signaling", "FOXO signaling" and "chromatin assembly during cell cycle", pointing to a miR-126-dependent core circuit shared between human Ph+ B-ALL and the mouse model. We started to validate the transcriptomic data at the protein level measuring phosphorylation status and subcellular localization of signaling hubs. Primary B-ALL from patient 2 showed increased phospho-STAT5 shortly after miR-126-OE, which returned to baseline within 6 days, suggesting adaptive responses counteracting a potentially deleterious excess of signaling in this disease. We hypothesize that a surge in signaling strength upon acute induction of miR-126-OE may have contributed to the apoptotic phenotype observed in patient 1. In keeping with this, constitutive miR-126-OE induced apoptosis in an additional 5 Ph+ B-ALL and, consequently, reduced engraftment in NSG mice, similar to knockdown of miR-126. Our results further consolidate an important role of miR-126 in human B-ALL, whereby a tight regulation of this miRNA appears critical for leukemia maintenance.