We investigated whether gamma-amino butyric acidergic (GABAergic) cell populations correlate positionally with specific Dlx-expressing histogenetic territories in an anamniote tetrapod, the frog Xenopus laevis. To that end, we cloned a fragment of Xenopus GAD67 gene (xGAD67, expressed in GABAergic neurons) and compared its expression with that of Distal-less-4 gene (xDll-4, ortholog of mouse Dlx2) in the forebrain at late larval and adult stages. In Xenopus, GABAergic neurons were densely concentrated in xDll-4-positive territories, such as the telencephalic subpallium, part of the hypothalamus, and ventral thalamus, where nearly all neurons expressed both genes. In contrast, the pallium of Xenopus generally contained dispersed neurons expressing xGAD67 or xDll-4, which may represent local circuit neurons. As in amniotes, these pallial interneurons may have been produced in the subpallium and migrated tangentially into the pallium during development. In Xenopus, the ventral division of the classic lateral pallium contained extremely few GABAergic cells and showed only low signal of the pallial gene Emx1, suggesting that it may represent the amphibian ventral pallium, homologous to that of amniotes. At caudal forebrain levels, a number of GABAergic neurons was observed in several areas (dorsal thalamus, pretectum), but no correlation to xDll-4 was observed there. The location of GABAergic neurons in the forebrain and their relation to the developmental regulatory genes Dll and Dlx were very similar in Xenopus and in amniotes. The close correlation in the expression of both genes in rostral forebrain regions supported the notion that Dll/Dlx are among the genes involved in the acquisition of the GABAergic phenotype.

To investigate the pallial organization and the exact location and extension of the ventral pallium in amphibians, we cloned a fragment of the homeobox XenopusTbr1 (xTbr1) gene and analyzed its expression compared with that of the genes xEomes (Tbr2) and xEmx1 in the telencephalon of the frog Xenopus laevis during embryonic and larval development. The expression of xEmx1 was also analyzed in the adult frog. We compared the expression patterns of these pallial marker genes with that of the subpallial gene xDistal-less-4 (xDll4). Our results indicate that the whole pallium of Xenopus expresses the T-box genes xTbr1 and xEomes (in proliferating cells and/or mantle) during embryonic and larval development, and the expression of these genes is topographically complementary to that of xDll4 in the subpallium. In addition to their massive expression in the pallium, both xTbr1 and xEomes are expressed in a few dispersed cells in the subpallium, which may represent immigrant cells of pallial origin, because these genes are not found in the subpallial proliferating cells. On the other hand, during development xEmx1 is expressed in a large part of the pallium (proliferating and postmitotic cells) except for an area adjacent to the pallio-subpallial boundary, where xEmx1 is observed only in some mantle cells. This pallial area poor in xEmx1 expression and poor in expression of the subpallial gene xDll4, but expressing the pallial marker genes xTbr1 and xEomes, appears to represent the amphibian ventral pallium, comparable to that described in other vertebrates (Puelles et al. [2000] J. Comp. Neurol. 424:409-438). In the adult frog, the ventral pallium appears to include the rostral part of the lateral amygdalar nucleus as well as a large part of the medial amygdalar nucleus (as defined by Marín et al. [1998] J. Comp. Neurol. 392:285-312). In contrast, the caudal part of the previously termed lateral amygdalar nucleus shows strong xEmx1 expression and may be a lateral pallial derivative. The possible homology of these amphibian amygdalar nuclei is discussed. Finally, expression of xTbr1, xEomes, and xEmx1 is observed in the mitral cell layer of the olfactory bulb from early developmental stages, further supporting that this structure is a pallial derivative.

To define at the molecular biological level the effects of thyroid hormone on brain development we have examined cDNA clones of brain mRNAs and identified several whose expression is altered in hypothyroid animals during the neonatal period. Clones were identified with probes prepared by subtractive or differential hybridization, and those corresponding to mRNAs altered in hypothyroidism were further studied by Northern blot analysis. Using RNA prepared from whole brains, no effect of hypothyroidism was found on the expression of the astroglial gene coding for glial fibrillary acidic protein. Among genes of neuronal expression, no significant alterations were found in the steady state levels of mRNAs coding for neuron-specific enolase, microtubule-associated protein-2, Tau, or nerve growth factor. N-CAM mRNA increased slightly in hypothyroid brains. In contrast a 2- to 3-fold decrease was found in the mRNA coding for a novel neuronal gene, RC3. This is the first neuronal gene known to be significantly altered at the mRNA level by thyroid hormone deprivation. The abundance of the mRNAs for the major myelin proteins proteolipid protein, myelin basic protein, and myelin-associated glycoprotein, expressed by oligodendrocytes, were also decreased in hypothyroid brains. Developmental studies on RC3 and myelin-associated glycoprotein expression indicated that the corresponding mRNAs accumulate in the brain of normal rats during the first 15-20 days of neonatal life. A similar accumulation occurred in hypothyroid brains, but at much reduced levels. The results demonstrate that thyroid hormone controls the steady state levels of particular mRNAs during brain development.

We have determined the concentration of thyroid hormone receptor binding sites in nuclear extracts derived from rat fetal organs throughout gestation and the postnatal period. Before day 14 of gestation nuclear extracts were obtained from whole fetuses. No receptor binding activity could be detected at day 12 of gestational age, and small amounts were detected at day 13 (maximum binding capacity less than 50 fmol/mg DNA). The receptor could be measured in pools of individual organs from day 14 (brain) or from day 16 (heart, liver, and lung) onwards. The order of analog binding affinity at 14 days was triiodothyroacetic acid = T3 greater than T4 greater than rT3, suggesting that at 14 days of fetal age the receptor has the same binding specificity as the receptor from mature tissues. In brain, the concentration of binding sites increased from 77 fmol/mg DNA at 14 days to 210 fmol/mg DNA at 17 days, remaining at this level until birth. Receptor concentration was identical whether the binding assays were performed on purified nuclei or nuclear extracts. There was no effect of maternofetal hypothyroidism on receptor concentration in the brain at 21 days of gestational age. Lung concentrations of receptor also remained constant during the fetal period. During the postnatal period, there was an increase in receptor concentration in brain and lung, with maximum levels at day 6. The pattern of receptor development in heart and liver was different, since its concentration increased progressively throughout the fetal and postnatal periods towards the levels found in adult rat tissues. The results suggest that the appearance of the thyroid hormone receptor coincides with that of the first fetal thyroid gland structures, but that it occurs much before thyroid function is fully established. As far as the receptor is concerned, fetal tissues have the potential to respond to thyroid hormone as early as the 13th day of gestational age.

The fate of the anterior neural ridge was studied by following the relative movements of simultaneous spot applications of DiI and DiO from stage 15 through stage 45. These dye movements were mapped onto the neuroepithelium of the developing brain whose shape was gleaned from whole-mount in situs to neural cell adhesion molecule and dissections of the developing nervous system. The result is a model of the cell movements that drive the morphogenesis of the forebrain. The midanterior ridge moves inside and drops down along the most anterior wall of the neural tube. It then pushes forward a bit, rotates ventrally during forebrain flexing, and gives rise to the chiasmatic ridge and anterior hypothalamus. The midanterior plate drops, forming the floor of the forebrain ventricle, and, keeping its place behind the ridge, it gives rise to the posterior hypothalamus or infundibulum. The midlateral anterior ridge slides into the lateral anterior wall of the neural tube and stretches laterally into the optic stalk and retina, and then rotates into a ventral position. The lateral anterior ridge converges to the most anterior part of the dorsal midline during neural tube closure, then rotates anteriorly, and gives rise to telencephalic structures. Whole-mount bromodeoxyuridine labeling at these stages showed that cell division is widespread and relatively uniform throughout the brain during the late neurula and early tailbud stages, but that during late tailbud stages cell division becomes restricted to specific proliferative zones. We conclude that the early morphogenesis of the brain is carried out largely by choreographed cell movements and that later morphogenesis depends on spatially restricted patterns of cell division.