During embryonic development, the cell fates are specified by the expression of lineage factors whose expression must be finely regulated to ensure proper organ formation. To get insight into the control of lineage factors expression, we used the vertebrate hindbrain (rhombencephalon) as a system model. During development, the hindbrain is subdivided into seven segments, termed rhombomeres and noted from r1 to r7. Each segment corresponds to a coherent cell lineage. In particular, specification of r3 and r5 lineages is controlled by the zinc-finger transcription factor Krox20. When Krox20 function is abolished, r3 and r5 cells are not properly specified, hence modified neuronal fate. In the present PhD work, we aimed at deciphering the processes that control the number of Krox20-expressing cells, i.e. the size of r3 and r5. We focused in particular on the transcriptional control of Krox20 expression by studying the activity of Krox20 cis-regulatory elements. Two elements, termed B and C, are responsible for the initiation of Krox20 expression; a third one, noted A, amplifies and prolongs it through an autoregulatory activity. We showed (i) that cells commit to the r3/r5 fate only if they activate the element A, (ii) that the element A functions as a bistable switch, determined by the level of Krox20 initiation. These results were obtained by the analysis of a computational model, constrained by quantitative data derived from experiments on zebrafish embryos. Our model is built at the cell level and implement molecular events occuring at the level of element A. It thus establishes the relationship between enhancer activity and tissue-scale patterning
