Oxytocinergic Cells of the Hypothalamic Paraventricular Nucleus Are Involved in Food Entrainment

In the brain several structures show rhythms coupled to the light/dark cycle including the paraventricular hypothalamic nucleus (PVN). Retrograde tracing studies indicate that the PVN is a target output from the suprachiasmatic nucleus (SCN) and they show a parallel phase-dependent induction of Fos protein, the product of the c-Fos gene, in response to a light pulse (Munch et al., 2002) and explants of this nucleus in culture do not show circadian rhythmicity (Abe et al., 2002). Overall the above results indicate that the activity of the PVN is rhythmic but is driven by the SCN. In the intact rat cells showing Fos expression in the PVN increase in numbers during the night in comparison to the day (Nunez et al., 1999). It is widely recognized that this time keeping mechanism generated from the SCN to other brain structures as the PVN, is present in several central and peripheral organs to ensure that bodily processes are carried out at the appropriate, optimal time of day or night (Kriegsfeld and Silver, 2006). However, when subjects are exposed to food for a few hours daily, this orderly hierarchy uncouples from the SCN. Animals develop an increase in locomotor behavior a few hours before food presentation, which is termed food anticipatory activity (FAA; Mistlberger, 1994). Other metabolic and hormonal parameters as well as central structures are also entrained by timing of food, even in the absence of the SCN (Caba and Mendoza, 2018). Food-intake then is a cue that elicits physiological responses, which act as entrainment stimuli for the brain and peripheral organs (Escobar et al., 2009; Mistlberger, 2011). However, despite considerable research effort, the internal entrainment stimuli are not well understood nor how central and peripheral oscillators are coupled (Mistlberger, 2011). In the present contribution we explored the oxytocinergic (OTergic) system of the hypothalamic supraoptic (SON) and PVN, the main sources of OT in the brain (Swanson and Kuypers, 1980) in relation to food-entrainment.

In contrast to the low activity of the PVN during the day (Nunez et al., 1999) this nucleus shows a sharp increase in Fos protein, after food presentation during the day in food entrained adult rats (Angeles-Castellanos et al., 2004). This result is interesting because it is possible that this increased activity of the PVN could be part of an afferent system that brings information about food intake; however, their phenotypical identity has not yet been explored. In present contribution we explored OTergic cells of both SON and the PVN 12 h before, and immediately before and after food ingestion by using the rabbit pup, which is considered a natural model of food restriction (Caba and Mendoza, 2018). Whether nursing occurs during the day or the night, pups show intense FAA before arrival of the mother (Caba et al., 2008). In the rabbit pup, metabolic, and physiological parameters such as corticosterone, and neural activity in some structures, shift their rhythm in relation to nursing time, similar to rodents where mealtime can act as a zeitgeber (Morgado et al., 2008, 2010; Caba and Mendoza, 2018). We found a differential activation of OTergic cells in different subregions of the PVN in relation to timing of feeding. This perhaps could be related to a possible main role of OT cells of the PVN both as an entraining signal and as an important pathway for the coordination between central and peripheral structures for the organized oscillation of the system that leads to the food entrainment phenomenon, which remains to be determined.

The main findings of the present study are: (1) There is an activation of OT neurons after feeding in the SON and PVNmv, (2) OT cells of the PVNc and in the dorsal part of PVNm shows activation before feeding, and moreover, this activation persists in fasted subjects at the previous scheduled time of nursing. The PVN is a complex neural structure that integrates essential endocrine and autonomic responses. In the rat it is comprised of several distinct morphological and neurochemical heterogeneous populations of magnocellular and parvocellular cells organized in 10 structural subdivisions (Koutcherov et al., 2000). Magnocellular regions of the SON and some regions of the PVN contain OT or vasopressin (AVP) producing cells that project to the posterior pituitary to release these peptides into the bloodstream (Swanson and Kuypers, 1980). However, parvocellular regions are more complex as they project to other forebrain regions as well as to the brainstem and spinal cord enabling them to influence the activity of the autonomic nervous system (Swanson and Kuypers, 1980). Moreover, their phenotypical identity is more diverse; they contain dopamine, somatostatin, encephalin or corticotrophin-releasing factor, besides OT or AVP (Koutcherov et al., 2000). The cytoarchitectonic subdivisions of the PVN are well characterized in rodents, particularly in the rat (Swanson and Kuypers, 1980), but not in others as in humans (Koutcherov et al., 2000). Also, in the rabbit this nucleus has a loose organization without any obvious subdivision (Schimchowitsch et al., 1989).

As shown in Figure 3C, the double label study indicates that most OT cells of the PVNmv are inactive before feeding but milk ingestion induces Fos protein in many of these cells. On the contrary OT cells from the PVNmd and PVNc show activation before the nursing time. This differential activation of OTergic cells in the different experimental conditions suggests a distinct function. By comparison with rat studies, the activation of OT cells in the PVNmv and the SON, is probably more related to an endocrine response to milk intake as food ingestion induces a massive release of OT into the bloodstream (Verbalis et al., 1996) to act peripherally to control metabolic homeostasis (Onaka and Takayanagi, 2019). OT improves insulin sensitivity, increases lipolysis, enhances glucose uptake and lipid utilization in adipose tissue and skeletal muscle and plasma OT levels are notably lower in obese individuals with diabetes (Spetter and Hallschmid, 2017; Ding et al., 2019). However, there is also release of OT in the hypothalamus in regions involved in food intake and energy expenditure as the dorsomedial, arcuate and ventromedial nuclei and the lateral hypothalamus, which contain oxytocin receptors (Onaka and Takayanagi, 2019). In general OT is considered an anorexigenic hormone that inhibits food intake and promotes satiety after food ingestion (Arletti et al., 1989; Olson et al., 1991; Onaka and Takayanagi, 2019).

Notwithstanding the importance of OT in food intake, satiety and homeostatic metabolism this hormone had not been involved in the phenomenon of synchronization by food. The concept of food entrainment includes an afferent pathway that entrains distributed central and peripheral oscillators that drives food anticipatory rhythms (Mistlberger, 2011). In this regard, food ingestion induces a large activation of both Fos protein (Angeles-Castellanos et al., 2004) and c-Fos mRNA (Poulin and Timofeeva, 2008) in the main body of the PVN in adult rats under food restriction. In the rabbit model we previously found a similar activation of cells in the main body of the PVN after their only daily period of milk ingestion, which were identified as OTergic cells (Caba et al., 2003), similar to present results. This OT activation was a consequence of food ingestion, as it was not observed in un-nursed pups. In considering that food seems to be the entraining cue for the phenomenon of food entraining in rats (Mistlberger, 2011) and the rabbit pup (Morgado et al., 2011) we propose that those OTergic cells of the main body of the PVN and SON, may play a main role as an entraining cue.

Additionally, our detailed observation of the PVN at different time points of feeding in nursed and un-nursed subjects led us to identify that two subpopulations of OT cells one in the dorsal part of the main body of PVN and those of PVNc, unlike OT cells of the PVNmv, are active before their only period of food ingestion, i.e., they “anticipate” feeding time. Moreover their activation is not permanent, as it is not observed 12 h after feeding although it is observed again at the previous scheduled feeding time even when food was omitted in un-nursed pups. This activation before feeding time agrees with a previous publication in rats. Poulin and Timofeeva (2008) reported an increase of c-Fos mRNA in the mediodorsal subdivision of the PVN just before food intake. In the corresponding figure (see Figure 6 in Poulin and Timofeeva, 2008) it is possible to identify that this region lies in the dorsal portion of the PVN, at a similar level identified as PVNmd in our results. We propose that the activation of these PVN cells that anticipate feeding time could be related to the peripheral metabolic changes observed in food entraining (Escobar et al., 2009) through indirect projections to peripheral organs. It is important to consider that PVN as well as SON are the main sources of AVP in the hypothalamus and future experiments are warranted to explore the potential participation of this hormone in food entrainment.

Classical studies demonstrate that PVN OT neurons besides the pituitary also projects to the spinal cord and brainstem regions as the dorsal vagal complex (Swanson and Kuypers, 1980; Sofroniew, 1985). Further studies of denervation of parasympathetic and sympathetic nerves combined with injection of pseudorabies virus (PRV), a transynaptic tracer, in the pancreas, revealed that second order neurons in several brain regions including the hypothalamus and cerebral cortex support a role of these central structures in the endocrine activity and cephalic response of the pancreas to food (Buijs et al., 2001). Furthermore, these authors, by using a similar strategy, also found preautonomic OTergic neurons in the posterior part of the PVN neurons that project to sympathetic or parasympathetic motor nuclei connected with the adrenal and liver, besides the pancreas (Buijs et al., 2003). In agreement, in vivo studies demonstrate that activation of sympathetic or parasympathetic innervation of the liver as well as pharmacological manipulation of preautonomic PVN neurons affects glucose homeostasis (O’Hare and Zsombok, 2016). These studies are very interesting in considering that in subjects under food restriction there is a shift in several hormonal and metabolic parameters to timing of feeding. Under this condition before food intake subjects show an increase in corticosterone, free fatty acids and glucagon and a decrease in glycogen and insulin, indicating they are in a catabolic state (Díaz-Muñoz et al., 2000). Moreover, authors also detected an oxidized condition of the liver before food access that suggests a preparatory process to the upcoming food intake (Díaz-Muñoz et al., 2000). In the rabbit pup we previously reported an increase in corticosterone and free fatty acids and a decrease in glycogen, which agree with the proposed catabolic state of subjects before food intake (Morgado et al., 2008, 2010). In considering the anticipatory changes in the liver and pancreas and the control of these organs by the sympathetic and parasympathetic system it is possible that the activated OT cells in PVNmd and PVNc of our present study could play an important role in the physiological preparation during FAA. In addition, in the rat lesions of the catecholaminergic input to the PVN suppress the anticipatory corticosterone peak in food entrained rats (Honma et al., 1992) which may indicate that the observed anticipatory activation of PVN cells could also be a response to a factor released in advance of mealtime. Then we propose that those OT subpopulations can be an important link between the brain and the peripheral organs, as part of the putative multi oscillatory mechanism of the FEO. The facts that in un-nursed pups these cells are active again (present results) together with metabolic and hormonal parameters at the scheduled previous time of nursing (Morgado et al., 2008, 2010), strongly reinforce this assumption. Finally in spite of the possible importance of the PVN cells in food entrainment, lesions of this nucleus indicates that the PVN it is not necessary for food anticipatory activity rhythms (Mistlberger and Rusak, 1988), similar to other brain nuclei (Mistlberger, 2011).

The SON could be more related to the homeostatic actions of OT at peripheral level in considering that this nucleus mainly projects to the posterior pituitary (Swanson and Kuypers, 1980; Sofroniew, 1985). Other hormones as corticosterone, ghrelin, leptin, and glucagon are entrained or seem to play a modulatory role in food entrainment (Patton and Mistlberger, 2013). On basis of present results we consider that is worth to explore OTergic cells in detail by discrete lesion or stimulation (Atasoy et al., 2012) of particular subpopulations of the PVN as they can contribute to the understanding of the food-entraining phenomenon. Additionally, food-entrainment implicates a daily reward mechanism that involves dopamine (Gallardo et al., 2014). On this regard it is interesting that oxytocin regulates feeding also through mechanisms in reward related areas as the nucleus accumbens, amygdala, bed nuclei of the stria terminalis and ventral tegmental area, which contain oxytocin receptors (Lawson et al., 2019).

OTergic cells of the SON and PVNmv may play a role as an entraining signal for synchronization by food. Besides discrete populations of OT cells in PVNmd and PVNc can be a link between neural and hormonal and metabolic changes through projections of the autonomic nervous system to prepare subjects during FAA to optimize food ingestion and use of nutrients during the short period of food presentation. Moreover, the persistence of activation of these OT cells at the time of the scheduled next nursing time in un-nursed subjects suggests a self-sustaining oscillation, which perhaps plays a main role in the multioscillatory system in control of food-entrained rhythms.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

The animal study was reviewed and approved by the Ethics Committee of Universidad Veracruzana.

MC and EM conceived and designed the experiments, and wrote the manuscript. EM, CH, MH, and MR-H performed the experiments. MC, EM, CH, and MH analyzed the data. MR-H, CH, and MH contributed with reagents, design, and elaboration of figures.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

This work was supported by the CONACYT Scholarships (#71818) to CH and (610438) to MH. We gratefully acknowledge Dr. Alyson S. Fleming and Dr. Roberto Meza-Romero for corrections and helpful comments.