The mouse Hox2.3 homeobox-containing gene: regulation in differentiating pluripotent stem cells and expression pattern in embryos.

Genomic and cDNA clones of the mouse Hox2.3 gene have been isolated. Expression of this gene was characterized in differentiating embryonal carcinoma (EC) and embryonic stem (ES) cells, and in the 13.5-day embryo. Hox2.3 is expressed at a very low level, if at all, in undifferentiated ES and EC cells. As previously reported for the Hox1.1 and Hox2.1 genes, differentiation of pluripotent stem cells induced by a nonchemical method is not accompanied by strong accumulation of Hox2.3 transcripts. Treatment of the stem cells with a chemical inducer like retinoic acid (RA), and also hexamethylenebisacetamide (HMBA), or 5-bromo-2'-deoxyuridine (BUdR), simultaneously accelerates differentiation and stimulates accumulation of Hox2.3 mRNA to high levels. Addition of RA several days after the cells have been induced to differentiate by a nonchemical method induces Hox2.3-transcript accumulation as well. For comparison, expression of the En-1 gene, which contains a homeobox belonging to a different class from that of the Antennapedia-related Hox1.1, Hox2.1, and Hox2.3 genes, was analyzed. The En-1 gene was found also to be sensitive to this regulation by chemical inducers of differentiation. It was observed that treatment of undifferentiated EC cells with the inhibitor of protein synthesis cycloheximide resulted in slight accumulation of Hox2.3 mRNA, suggesting the involvement of a short-lived protein in keeping the level of homeobox-gene transcription low in EC cells. The highest level of Hox2.3 transcripts in 13.5-day embryos in vivo was observed in the spinal cord. Comparison with the expression pattern of three other homeobox genes revealed overlapping gradients of mRNA along the longitudinal brain-spinal-cord axis. An important question is that of the molecular basis for such a spatially restricted accumulation of homeobox transcripts. Hox2.3 is expressed at a much lower level in rat and mouse embryonic midbrain than in spinal cord in vivo. We have shown that addition of RA to primary cultures of cells from rat embryo mesencephalon leads to strong accumulation of Hox2.3 mRNA. A possible interpretation is that RA mimics one or more spatially restricted effectors, accounting for the local accumulation of Hox2.3 transcripts in the embryonic central nervous system. Control of Hox2.3 gene expression in vivo may obey some similar mechanisms as in chemically stimulated EC and ES cells in vitro.