Figure 1: Fetuses with abnormal axial skeleton--the ribs and vertebrae--due to exogenous addition of retinoic acid to pregnant mice. (A) represents the wild-type. Vertebrae are obviously missing in (B) and (C).
Teratogenic doses of retinoic acid has been found to activate some Hox genes not normally expressed in groups of cells along the anterior-posterior axis. An experiment by Kessel and Gruss (1991) demonstrated that the addition of retinoic acid affected vertebral development by leading to less vertebrae being expressed and posterioralization of which vertebrae were expressed.
Characterized by a 183 base pair motif that encodes a DNA-binding domain, the Hox genes have been found to function in vertebrate axes specification. Partially overlapping domains of the expression of the Hox genes are thought to create a Hox code, where Hox genes are expressed at different axial levels in distinct combinations. The Hox code hypothesis says that certain combinations of Hox genes specify a specific region along the anterior-posterior axis. Gene knockout experiments and retinoic acid-induced homeosis support this, as experiments adding retinoic acid exogenously changes not only changed the patterning of homeotic gene expression along the anterior-posterior axis but also led to abnormal differentiation of axial structures. In summary, without Hox genes, the organ will lack specific organs along the anterior-posterior axis or, in the case of retinoic acid induced homeosis experiments, structures get respecified to another region along the axis. The evidence that a Hox code determines patterning includes that there is a need for different sets of Hox genes to have complete segmentation of the regions of the axis. A set of paralogous genes is also responsible for organs to form fully (Figure 2).
Figure 2: The four mammalian Hox clusters A-D. Each box represents a Hox genes. A box with a question mark represents genes which have not yet been discovered in vertebrates. Note that the numbering of the Hox genes assume that these genes will be found eventually (from Gudas, 1994).
In itself, trans retinoic acid (tRA) may not be a morphogen, but instead may interact with other induction signals including sonic hedgehog, Bone Morphogenetic Protein-2 (BMP 2), and Fibroblast Growth Factor-4 (FGF4) . Retinoids are thought to function as morphogens during anterior-posterior patterning in vivo (Durston et al., 1996). The beginnings of much of the anterior-posterior specification in Xenopus is thought to be established during gastrulation. The mesoderm in the dorsal blastopore lip of the gastrula functions as the organizer by emitting certain signals at sequential steps which leads to the development of different structures of the body along the anterior-posterior axis of the body. The addition of tRA has been found to disrupt the activity of the organizer.
Retinoic acid affects the development of head structures in a concentration-dependent manner and can alter tail structures as well. Resembling a neural transformation signal, tRA affects central nervous system patterning by posterioralizing induced neural tissue, but in itself, is not a neural inducer (Durston et al., 1996). Adding tRA to whole Xenopus embryos leads to increases in volume in the hindbrain and spinal cord with a corresponding decrease in volume of the forebrain. Treatment with tRA late in gastrulation will lead to the expression of posterior neural markers in the presumptive forebrain regions of the embryo. Treatment early in gastrulation, however, compacts or deletes the presumptive hindbrain and also disrupts hindbrain segmentation. Along with the Hox code hypothesis, this evidence shows that retinoic acid functions affects the specification of the structures along the anterior-posterior axis of the Central Nervous System (CNS). Axial mesoderm polarity is also posterioralized by retinoic acid. The specific teratogenic function of retinoic acid is thought to be conserved in different vertebrate species. Different organizer regions use conserved signals, one of which may be a retinoid.
Evidence exists which supports the premise that tRA is a morphogen because it is available endogenously in the embryo. Because the retinoid posterioralizes embryos and induces Hox gene expression where the genes are normally not expressed, it is thought that tRA is localized in the posterior region of the dorsal blastopore lip. At least four retinoids have been found to be sequestered in the cells of this area and synthesized locally in developing embryos.