Loss of miR-146a is implicated in severe graft-versus-host-disease (GvHD) – a rate-limiting toxicity of allogeneic hematopoietic stem cell transplantation. We screened a panel of chemically modified cholesterol-conjugated miR-146a mimics in order to identify a clinically relevant chemical scaffold for GvHD therapy.

We tested the impact of the metabolic stability of the antisense strand as well and length and degree of complementarity of the sense strand of miR146a. A fully complementary sense strand impaired the silencing efficacy on an siRNA-like target reporter 147-fold (p<0.0001). However, a shorter sense strand left silencing efficacy largely unaffected in the siRNA-like target reporter assay. Silencing efficacy of the miR146a target IRAK1 was impaired 8.6-fold (p=0.02) by fully complementary sense strands but not by shorter sense strands. Silencing was completely abolished when combining the above sense strand modifications (p<0.0001). Interestingly, in an in vitro model of GvHD, fully complementary sense strands rather improved the efficacy of inhibiting T cell proliferation in the context of both longer and shorter sense strands (1.7-fold and 1.5-fold, respectively, p<0.0001). The metabolic stability of the antisense strand did not significantly affect efficacy of miRNA mimics in either assay. We were able to inhibit T cell proliferation upon allogeneic stimulus up to 80% with an IC50=2.8 µM. Furthermore, we showed that chemically synthesized miR-146a mimics do not impair the T-cell-mediated anti-leukemia effect in vitro.

Collectively, we demonstrate that fully chemically modified miRNA mimics may be useful to treat GvHD and that the duplex structure of chemically synthesized miRNA mimics predicts miRNA efficacy.