Molecular basis of positional memory in limb regeneration [RNA-seq]

Amputation of a salamander limb triggers anterior and posterior connective tissue cells to form distinct signalling centres that together fuel regeneration. Anterior and posterior identities are established during development and are thought to persist lifelong in the form of positional memory. However, neither the molecular basis of positional memory nor whether positional memory can be altered is known. Here, we identify a positive feedback loop responsible for posterior identity in the axolotl (Ambystoma mexicanum) limb. Posterior cells express residual Hand2 transcription factor from development, which primes them to form a Shh signalling centre after limb amputation. During regeneration, Shh signalling is also upstream of Hand2 expression. After regeneration, Shh is shut down but Hand2 is sustained, safeguarding posterior memory. We exploited this regeneration circuitry to convert anterior cells to a posterior cell memory state. Transient exposure of anterior cells to Shh during regeneration kick-started an ectopic Hand2-Shh loop, leading to stable Hand2 expression and a lasting competence to express Shh. Our results implicate positive feedback in the stability of positional memory and reveal that positional memory is more easily reprogrammed in one direction (anterior to posterior) than the other. Modifying positional memory in regenerative cells changes their signalling outputs, which has implications for tissue engineering. Overall design: We wanted to identify anterior-posterior identity transitions induced in axolotl cells after transplantation or after misexpression of Hand2. This GEO accession includes RNA sequencing data from 6 sample types. We sequenced FACS purified anterior and posterior blastema cells from regenerating axolotl limbs and prepared libraries with higher RNA input (45 ng per sample, 4 replicates each) or lower RNA input (4.5 ng, 4 replicates each). We also sequenced anterior cells transplanted to the anterior or posterior of a host axolotl limb to see what gene expression changes occur (4.5 ng input, 2 replicates each). We also sequenced anterior blastema cells misexpressing exogenous mCherry-Hand2 or, as a control, mCherry only (0.7 ng input, 3 and 2 replicates respectively).