Gene Mutations and microRNA Signature in Myelodysplastic Syndromes during Azacitidine and Lenalidomide Therapy

Background and Rationale. Several clinical studies demonstrated the efficacy and safety of combining Azacitidine (AZA) with Lenalidomide (LEN) in Myelodysplastic Syndromes (MDS), but the molecular implications of this therapy are still under investigation. Here we analyzed the effect of AZA+LEN therapy on MDS, particularly focusing on gene mutations as well as microRNA profiling.

Patients and Methods. The study included 44 patients, diagnosed with higher risk MDS (IPSS High or Intermediate-2), who received AZA (75 mg/m2/day, days 1-5, sc) and LEN (10 mg/day, days 1-21 or 6-21, orally) every 4 weeks. Patients were considered clinically evaluable after at least 6 cycles of treatment. Response was assessed according to the IWG response criteria (Cheson 2006). Molecular analyses were performed on mononuclear cells extracted at baseline and during therapy (after cycles 4th and 8th). Gene mutations were investigated using a NGS myeloid gene panel (31 genes), whereas the expression of microRNAs was assessed using a 4.0 miRNA Affymetrix array.

Results. Of the 44 patients, 34 were clinically evaluable for response, with an ORR of 76.5% (26/34 cases). Molecular analyses were performed on these 34 cases. For gene mutations, 30 paired samples (pre- and post-treatment) were analyzed: 3 patients showed no mutations neither at baseline nor during therapy, all other patients had at least one mutation. The most frequent mutations were ASXL1 (14 cases=47%), TET2 (11 cases=37%), RUNX1 (8 cases=27%) and SRSF2 (5 cases=17%). All samples with a decreasing variant allele frequency (VAF) were responders, while all the non-responders had at least one gene mutation and none showed a decreasing VAF.

Kaplan-Meier analyses revealed that the presence of either CEBPA or CBL mutations, detected respectively in 3 and 2 patients, were associated with a decreased OS (28 vs 14 months and 28 vs 19 months, respectively; p<0.05) and duration of the therapy (28 vs 10 months and 28 vs 15.5 months, respectively; p<0,05). Only CEBPA mutation was significantly associated with the duration of response (27 vs 6 months; p<0.05), while CBL mutation was close to significant (27 vs 9 months, p=0.08). As for the other gene mutations, some genes were significantly associated only with duration of therapy (NRAS, RUNX1, TP53, SF3B1, U2AF1; p<0.05) and others were significantly associated only with the duration of response (TP53, IDH2; p<0.05).

For microRNA profiling, to date the analyses were performed on 12 patients, in order to collect paired samples at baseline and after both the 4th and the 8th cycle of therapy. As compared to baseline, responders and non-responders showed a different molecular pattern: at the 4th cycle we had a cluster of 43 microRNAs up- or down-regulated and at the 8th cycle the cluster included 19 microRNAs. Interestingly, 4 of these 12 patients were responders at the 4th cycle and lost the response at the 8th cycle. In this group of patients, 8 microRNAs showed a statistically significant difference between the 4th and the 8th cycle, with 5 up-regulated (e.g. miR-4767) and 3 down-regulated microRNAs (e.g. mir-424).

Conclusions. Our results show that AZA+LEN therapy in high-risk MDS patients can induce a favourable clinical response and that the molecular mutation profiling could be important to better understand the effect of the therapy. In addition, a specific mini-cluster of differentially expressed microRNAs could be associated with the loss of response. Further studies are needed to confirm all these data, and to identify the specific targets for these microRNAs, as some of them could target or be strictly associated with important molecules in MDS pathogenesis, such as those of the nuclear inositide signalling pathways.