Our discovery that the SUMO chromatin landscape differs greatly when comparing differentiated cells (mouse embryo fibroblasts, MEFs) with stem cells (mESCs), led us to hypothesize similarly important differences in the composition of the SUMOylome between these cell types. Using a sensitive mass-spectroscopy approach, we identified the SUMOylated proteins in MEFs and found these to be largely comprised of proteins functioning in a variety of "normal" cellular processes that regulate MEF-specific gene expression. By contrast, mESCs displayed a highly interconnected network of SUMO targets involved in maintaining densely-compacted chromatin ("heterochromatin") usually associated with inactive (repressed) genes. These results indicate that SUMOylation contributes to the maintenance of the fragile stem cell state by globally repressing the activity of genes associated with cell differentiation, or even de-differentiation (reversion) to an earlier ("totipotent") cell state and further imply a massive rewiring of the cellular SUMOylome that accompanies changes in cell fate.
A most unexpected discovery was made when we treated mESCs at two intervals with a pharmacological SUMOylation inhibitor. Here, we found these cells to aggregate into spheroidal structures that self-organized into gastrula-like, patterned assemblages which, in turn, developed into "embryo-like structures" (ELSs) when cultured in a specially-designed droplet microfluidic device (see Figure). These structures possess a marked anterior-posterior axis and to contain properly positioned anterior neuronal cell types and paraxial mesoderm segmented into somite-like structures, thus recapitulating salient features of 8 day-old natural embryos. While these ELSs remain far from perfect in lacking, for example, proper dorso-ventral patterning and primordial germ cell clusters, these results indicate that pulsed SUMOylation inhibition, i.e. by only targeting an epigenetic regulator, can drive a single cell type, ESCs, into multiple, "self-organized" developmental trajectories that go well beyond the simple transitions of one cell type into another seen previously.
To extend the above principles to a tissue context -- muscle differentiation and regeneration after injury, our in vitro (ex vivo) and in vivo experiment have yielded preliminary results indicating that SUMOylation-compromised conditions yield higher numbers of satellite cells and that in a chronic injury model, mice from a SUMOylation-compromised genetic background display fewer muscular lesions and larger muscle fibers than control animals. While these results need to be confirmed, they tend to indicate that lowering SUMOylation levels may increase muscle regenerative capacity.
