The distribution of Dlx1-2 and glutamic acid decarboxylase in the embryonic and adult hypothalamus reveals three differentiated LHA subdivisions in rodents.

The lateral hypothalamus (LHA) is still a poorly understood brain region. Based on published Dlx and Gad gene expression patterns in the embryonic and adult hypothalamus respectively, three large areas are identified in the LHA. A central tuberal LHA region is already well described as it contains neurons producing the peptides melanin-concentrating hormone or hypocretin. This region is rich in GABAergic neurons and is specified by Dlx gene expression in the rodent embryo. Rostrally and caudally bordering the tuberal LHA, two Dlx-GAD-GABA poor regions are then easily delineated. The three regions show different organizational schema. The tuberal region is reticularly organized, connected with the cerebral cortex and the spinal cord, and its embryonic development occurs along the tractus postopticus. The region anterior to it is associated with the stria medullaris in both embryonic and adult subjects. The posterior LHA region is made of differentiated nuclei and includes the subthalamic nucleus. Therefore, the LHA is divided into three distinct parts: in addition to the well-known tuberal LHA, caudal and anterior LHA regions exist that have specific anatomical and functional characteristics. The hypothalamus is made up of several dozens of nuclei or areas that are more or less well differentiated and whose boundaries and arrangements are drawn differently according to authors and atlases (Allen Institute, 2004; Paxinos and Franklin, 2019; Paxinos and Watson, 2013; Swanson, 2004). The dominant hypothesis for more than 50 years is that these structures are distributed within three antero-posterior areas (anterior, tuberal, posterior) and more or less three longitudinal zones (lateral, medial and periventricular) (Fig. 1). In addition to these regions, several adjacent territories are often associated to the hypothalamus. The preoptic area is functionally related to the hypothalamus, but it is better seen as a telencephalic structure based on developmental data (Croizier et al., 2015; Puelles and Rubenstein, 2015). Lately, the zona incerta and the subthalamic nucleus (STN) have also been associated to the hypothalamus on the basis of their connections and development for the STN (Altman and Bayer, 1986; Barbier and Risold, 2021; Swaab et al., 2021). However, the zona incerta is still included in the 'pre-thalamus' or "ventral thalamus" in the embryo (Puelles and Rubenstein, 2015). Thus, the boundaries of the hypothalamus remain blurred around what we can call a 'core' made of the anterior to posterior regions (Brooks, 1988). In addition, unlike other large brain regions that are characterized early on by a molecular signature, i.e. by the embryonic expression of specific molecular markers, data illustrating the distribution of dozens of transcription factors involved in brain patterning and cell lineage specification confirmed the extremely heterogeneous and mosaic nature of the anterior and posterior regions of the hypothalamus (Alvarez-Bolado, 2019; Puelles et al., 2013; Puelles and Rubenstein, 2015). The rich nuclear organization of the medial and periventricular zones of the hypothalamus is consistent with the mosaic expression of developmental genes. The LHA, however, is often perceived as much more homogeneous in its cytoarchitectural organization. At the same time, there is little information regarding the expression of developmental genes in the anterior and posterior territories of the LHA. Most studies focus on the tuberal LHA which expresses many of these genes. Admittedly, even in the adult hypothalamus, the internal boundaries of the LHA are difficult to identify and the same is true in the embryo. Developmental data alone are insufficient to achieve a better understanding of the LHA anatomical organization and for this region as for medial and periventricular zones, a coherence must be established between development and adult anatomical organization. Among the most useful neurochemical markers to identify large regions of the forebrain, those involved in the identification of GABAergic and glutamatergic neurons have proven to be particularly efficient. Indeed, GABAergic neurons are not ubiquitously distributed. Large regions of the forebrain are rich in such cells, including the basal telencephalon, but others contain few or no GABAergic cells and are rich in glutamatergic neurons instead (for example the dorsal thalamus that is free of GABA-neurons in rodents). The same applies for the hypothalamus: several structures of the hypothalamus are free of GABAergic neurons, as, for example, the mammillary nuclei (Hahn et al., 2019). Recently, we also identified a GABA-poor posterior LHA territory that includes the (STN), and is localized caudal to the GABA-rich tuberal LHA (Barbier et al., 2020; Barbier and Risold, 2021; Chometton et al., 2016b). Therefore, the LHA seems partitioned into GABA-rich/GABA-poor regions. However, to define or confirm distinct neuroanatomical entities, these regions must have a specific embryological origin, and show specific hodological patterns and functions. Hence, the purpose of this short review is to identify divisions of the LHA based on developmental and neurochemical criteria. Such an analysis seems to us relevant in order to allow later functional studies on regions whose boundaries will be based on objective criteria.