In the early 1980s, we had observed that neuronal polarity varies depending on the respective topological origins of co‐cultured neurons and non‐neuronal cells (astrocytes). Following publication of these findings in 1984 (Denis‐Donini et al , 1984), we speculated that homeoprotein transcription factors (HPs) may regulate neuronal geometry, in analogy with the well‐established role of homeogenes (encoding HPs) in the regulation of organ shape in accordance with their position along the body anterior–posterior axis. In other words, we hypothesized that common mechanisms might contribute both to cell and to organ shapes, a pretty daring hypothesis at the time. To test it, we took advantage of the transient disruption of neuronal membranes during tissue dissociation, which allows for the passive internalization of proteins. Based on the high conservation of HDs, we anticipated, and indeed verified experimentally, that the HD Antennapedia could chase a variety of endogenous HPs off their cognate genomic binding sites. Changes in cell‐shape did happen thus validating the hypothesis that HPs could regulate neuronal shape. However, to rule out that this result was an artifact of the excess of HD added during tissue dissociation, the HD was added into the culture medium after cell plating. Indeed, the same morphogenetic changes took place, suggesting an experimental artifact. In a “just in case” experiment, we tagged the HD with fluorescein isothiocyanate to follow its distribution. This led to the rather astonishing discovery that the HD added to the culture medium was internalized by live, undisrupted neurons. We established that the internalization required the conserved 3rd helix of the homeodomain. Furthermore, this surprising finding turned out to be a general feature and we could show that a number of HPs signal directly between cells in vitro and in vivo by transferring as proteins between the cells (Fig 1). Although only about a dozen of …