Ruxolitinib, a Janus Kinase (JAK)-1 and -2 Inhibitor Modulates Microrna (miR) Levels in Patients with Myelofibrosis (MF): miR Levels Might Predict MF Treatment Outcome

Abstract 1744

Myeloproliferative neoplasms (MPN) are a class of stem cell–derived hematologic malignancies, characterized by an expansion of one or more myeloid lineages with resulting bone marrow (BM) hypercellularity. A gain-of-function mutation in JAK2, detected in most patients with MPN, induces constitutive activation of JAK2, stimulation of downstream signaling pathways, and activation of several cytokine and growth factor genes. Recent phase I/II clinical trial revealed that ruxolitinib (INCB018424), an oral inhibitor of JAK1 and JAK2, induced marked and durable clinical benefits in patients with primary or secondary MF (Verstovsek S et al. N Engl J Med 12:1117, 2010). Although clinically effective, ruxolitinib induced a modest (13%) suppression of the JAK2V617F allele burden after 12 cycles of therapy and the response did not correlate with the presence of JAK2V617 mutation. Therefore other, yet unknown, mechanisms might be operative in attaining these clinical benefits.

Alterations in gene and protein levels affect miRs levels, and miRs have been shown to regulate protein levels, thereby playing a role in cancer pathogenesis (Calin G, et al. PNAS 99:15524, 2002; Merritt MW et al. N Engl J Med 25:2641, 2008; Munker R, et al, Clin Sci 121:141, 2011). Therefore, in MF patients we investigated changes in the levels of RNase III enzymes Dicer and Drosha, which are the two main regulators of miR biogenesis. Deregulation of their expression has been indicative of possible miR alterations in various cancers. We also investigated expression of several miRs (h-miR-16, 21, 29a, 29b1, 29b2, 29c, 155, 181a, and 451). Bone marrow mononuclear cells (MNC) from six best responders in the ruxolitinib study, and MNC from six worst responders were collected (at baseline, one year and two years on therapy), plus six normal control bone marrow MNC, and used for this study. Using qRT-PCR we found that the levels of miR-16, -21, -29 (a, b1, b2, c), -155, -188a and -451 were significantly higher in MF than in normal controls. MiR-155 levels were significantly higher at baseline in best responders than worst responders (P = 0.024). However, after two years of treatment they significantly declined (P = 0.055). The expression of miR-188a was also significantly higher in the worst responders at baseline, as compared to the best responders, but stayed elevated during two years of therapy (P = 0.0021). The levels of RNase III enzymes Dicer and Drosha were significantly lower at baseline in MF MNC as compare to normal MNC. However, after two years of treatment, Dicer level significantly increased in the best responders but not in the worst responders (P = 0.0001). In conclusion, ruxolitinib treatment modulates miR levels and associated enzymes; miR profiling might be useful in predicting response to ruxolitinib.