Extracellular vesicles (EVs) from therapeutic cells contain a wealth of non-coding RNA species, which may target one or more disease-relevant pathways. In particular, tRNA-derived small non-coding RNAs (tsRNAs) are highly enriched in EVs from cardiosphere-derived cells (CDCs). Prior mechanistic studies showed that macrophages are not only first responders to EVs and their contents, but also robust reporters of bioactivity. Using macrophages to screen for bioactive tsRNAs, we identified a 5’ tRNA half, tREX1, with favorable effects on macrophage gene expression. In a mouse model of Duchenne muscular dystrophy (DMD), tREX1 produced macrophage-dependent improvements in left ventricular ejection fraction and hind limb peak isometric torque, which were associated with decreased cardiac and muscle fibrosis. Being chemically unmodified, tREX1 is susceptible to degradation by RNases, which likely decreases its potency. Therefore, we sought to optimize our therapeutic candidate, tREX1, by selective chemical modification, with a view of enhancing its potency. Ten new chemical entities (NCEs) were created using phosphorothioate, 2’-O-methyl, and locked nucleic acid modifications. The majority of these NCEs showed dramatically enhanced macrophage bioactivity and RNase resistance, without compromising cell viability. Careful study of the structure-activity relationships indicated that the phosphorothioate backbone has intrinsic, RNase-independent bioactivity that synergizes with tREX1’s sequence. Our lead candidate, TT1, a prototype of a new class of RNA drugs called exomers, has now advanced in the developmental pipeline to preclinical studies in mouse models of DMD. Continued efforts seek to use our lead optimization platform to discover additional exomers that may become next-generation RNA therapeutics for DMD.