Gene dosage effects on hypothalamic visceromotor cell types (Commentary on Duplan et al.)


The paraventricular nucleus of the hypothalamus (PVN) is a mosaic of anatomically and chemically specified cell types, in which each of the major avenues by which the brain regulates visceromotor function is represented amply, and in a topographically ordered manner (Swanson & Sawchenko, 1983). These include (i) magnocellular neurosecretory neurons that synthesize the peptide hormones, oxytocin (OT) and vasopressin (AVP), for release into the general circulation via the posterior pituitary, (ii) parvicellular neurosecretory cells that serve as the central arms of the hypothalamo-pituitary-adrenal and -thyroid axes, which are broadly involved in adaptations to stress and metabolic challenges, respectively, and (iii) neurons that project within the CNS to cell groups involved in central autonomic control, including vagal and sympathetic preganglionic neurons and the nucleus of the solitary tract (NTS), the principal CNS recipient of primary visceral sensory information. This latter, ‘pre-autonomic,’ contingent is not known to possess a distinctive neurochemical signature; instead, small subsets of it share phenotypes with adjoining neurosecretory populations, with OT, AVP and corticotropin-releasing factor predominating. As these output neuron classes may be called into play in various combinations to promote adaptation to specific physiologic conditions, the PVN has become an obligatory focus for studying integrative aspects of hypothalamic function.

A key factor in the specification of PVN cellular phenotypes during development is Sim1, a basic helix-loop-helix-PAS nuclear transcription factor identified by homology with a Drosophila gene, single-minded, which regulates CNS midline development. Murine Sim1 is required for the development of virtually all neurons in the PVN and the related supraoptic nucleus (Michaud et al., 1998), a pure magnocellular neurosecretory cell group. Whereas sim1 null mice die shortly after birth, heterozygous sim1 mutants are viable, and develop early onset hyperphagia and obesity (Michaud et al., 2001). Sim1 is, in fact, one of a handful of genes linked with monogenic human obesity.

In this issue of EJN, Duplan and colleagues (2009) further explore the developmental and functional consequences of sim1 haploinsufficiency, finding that two major classes of PVN visceromotor neurons are compromised. They report that sim1+/− mice display substantially reduced complements of AVP- and, particularly, OT-immunoreactive cells, and reduced mRNA expression, while major parvicellular neurosecretory PVN phenotypes are ostensibly unaffected (cf. (Kublaoui et al., 2008). Sim1 heterozygotes are as capable as wild type mice of defending plasma osmolality under milder dehydration conditions (substituting hypertonic saline for drinking water, but not more strenuous ones, presumably as a result of diminished natriuretic and water-retaining capabilities deriving from partial loss (or dysfunction) of magnocellular OT and AVP populations. The effects of reduced sim1 gene dosage were not limited to magnocellular cells, as heterozygotes also displayed a >70% reduction in PVN neurons labeled following retrograde tracer injections in a major target of autonomic-related PVN outputs, the dorsal vagal complex (NTS and dorsal motor nucleus of the vagus).

These new findings direct particular attention to the role of Sim1 in the development of preautonomic PVN outputs, their potential role in the obese phenotype of sim1+/− mice, and more generally in the control of food intake and obesity. The projection to the dorsal vagal complex, and particularly its OT component, has been implicated in feeding inhibition, by modulating the sensitivity of the NTS to visceral sensory cues (Blevins et al., 2004), including the anorexia associated with chronic dehydration (Rinaman et al., 2005). This may well relate to the new finding that sim1+/− mice eat less than wild type counterparts under dehydration conditions that similarly affect plasma osmolality in the two groups. In this light, recent evidence that central administration of OT can normalize food intake and weight gain of sim1+/− mice (Kublaoui et al., 2008) supports an involvement of such a mechanism in the obese phenotype associated with sim1 haploinsufficiency. Additional contributors remain to be identified. Only fragmentary information is available as to the extent to which the phenotypic diversity that characterizes autonomic-related PVN projections may reflect a degree of chemical coding within them. In addition, because of the extensive degree of axon collateralization within these pathways (Swanson & Kuypers, 1980), the involvement of other targets, notably vagal and sympathetic preganglionic neurons, remains to be evaluated.

Interpreting the effects of reduced sim1 dosage by germ line manipulation is complicated by the findings that viral vector-mediated modulation of sim1 expression in the PVN can increase or decrease food intake in mature wild type mice (Yang et al., 2006). As noted by Duplan et al. (2009), conditional gene disruption in postnatal animals will be required to isolate the postnatal role(s) of Sim1, whose transcriptional target genes are not well characterized. Similar approaches, likely exploiting axonal transport methods to achieve differential targeting of functionally differentiated PVN populations that may share neurochemical phenotypes, will be needed to parse the roles of paraventriculo-autonomic projections in visceral sensory and motor functions involved in regulating appetitive behavior and metabolism.