Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2018 Jan 4.
Published in final edited form as: Neuron. 2016 Dec 15;93(1):57–65. doi: 10.1016/j.neuron.2016.11.021

Bidirectional Anticipation of Future Osmotic Challenges by Vasopressin Neurons

Yael Mandelblat-Cerf 1, Angela Kim 1,2, Christian R Burgess 1, Siva Subramanian 1, Bakhos A Tannous 2,3, Bradford B Lowell 1,2,*, Mark L Andermann 1,2,4,*
PMCID: PMC5215952  NIHMSID: NIHMS836099  PMID: 27989461

SUMMARY

Ingestion of water and food are major hypo- and hyperosmotic challenges. To protect the body from osmotic stress, posterior pituitary-projecting, vasopressin-secreting neurons (VPpp neurons) counter osmotic perturbations by altering their release of vasopressin, which controls renal water excretion. Vasopressin levels begin to fall within minutes of water consumption, even prior to changes in blood osmolality. To ascertain the precise temporal dynamics by which water or food ingestion affect VPpp neuron activity, we directly recorded the spiking and calcium activity of genetically defined VPpp neurons. In states of elevated osmolality, water availability rapidly decreased VPpp neuron activity within seconds, beginning prior to water ingestion, upon presentation of water-predicting cues. In contrast, food availability following food restriction rapidly increased VPpp neuron activity within seconds, but only following feeding onset. These rapid and distinct changes in activity during drinking and feeding suggest diverse neural mechanisms underlying anticipatory regulation of VPpp neurons.

In Brief

Using electrophysiological and optical methods, MandelblatCerf and colleagues demonstrate that ingestion of water or food leads to rapid, presystemic decreases or increases in activity of vasopressin neurons, respectively. Surprisingly, learned sensory cues predicting water, but not food, also modulated activity.

INTRODUCTION

Learned autonomic and endocrine responses prior to ingestion can help mitigate anticipated disturbances to body homeostasis (Pavlov, 1902; Powley, 1977; Woods, 1991). Neuroendocrine motor neurons (Harris, 1951; Watts, 2015), which release specific hormones into the bloodstream in an activity-dependent manner, also likely play important roles in countering the effects of ingestion on internal homeostasis (Stricker and Hoffmann, 2007; Stricker et al., 2002). However, the extent to which changes in activity in these neurons predict and/or follow the onset of ingestive behavior remains unclear.

Pituitary-projecting vasopressin (VPpp) neurons in the supraoptic nucleus (SON) and the paraventricular nucleus (PVH) of the hypothalamus comprise an important class of neuroendocrine neurons (Harris, 1951; Watts, 2015). VPpp neurons react to slow increases in plasma osmolality by releasing the anti-diuretic hormone vasopressin, which acts on the kidney to decrease water excretion (Bourque, 2008; Nielsen et al., 1995; Sands et al., 2011; Watts, 2015). Previous studies observed decreases in blood vasopressin levels within minutes following water consumption, preceding and hence “anticipating” changes in plasma osmolality (Baertschi and Pence, 1995; Choi-Kwon et al., 1990; Geelen et al., 1984; Huang et al., 2000; Stricker and Hoffmann, 2007). However, such measurements cannot reveal whether VPpp neurons can adjust their activity within seconds, during the approach and/or initial ingestion of water. While electrical recordings from putative VPpp neurons have been conducted in rats (e.g., Poulain and Wakerley, 1982), the use of anesthesia precluded the study of ingestive behaviors. Anticipatory changes in firing were reported in SON neurons in awake non-human primates (Arnauld and du Pont, 1982), yet these data were limited by recording yield and by challenges in accurately identifying VPpp neurons.

Here, using optetrode electrophysiology and fiber photometry in awake mice (see Cohen et al., 2012; Garfield et al., 2016), we demonstrate diverse anticipatory changes in activity of genetically defined VPpp neurons within seconds prior to and following the onset of ingestive behaviors.

RESULTS

Anatomical Tracing of Pituitary-Projecting VP Neurons in Avpires−Cre/+ Mice

Previous studies in rats showed that the vast majority of SON-VP neurons, and a subset of PVH-VP neurons, project to the posterior pituitary (Armstrong, 1995; Bourque, 2008; Hatton, 1990; Leng et al., 1999). Using multiple anatomical tracing techniques in Avpires−Cre/+ mice (Pei et al., 2014), we confirmed that mouse VPpp neurons exhibited similar anatomical properties and that SON VPpp neurons did not send substantial collaterals to other brain areas (see Figure 1A, Figure S1, and Experimental Procedures).

Figure 1. Water Cues and Drinking Both Cause Rapid Drops in Spiking Activity of Individual VPpp Neurons.

Figure 1

Open in a new tab

(A) Selective expression of ChR2-mCherry in SON vasopressin neurons that project to the posterior pituitary (VPpp).

(B and C) Spiking activity for two example VPpp neurons. Both neurons demonstrated a rapid decrease in firing within seconds of presentation of the lickspout (green vertical dashed line), and again at onset of water availability (black vertical dashed line). Vertical ticks indicate licking prior to (yellow) and during (orange) water consumption. Significant decreases in firing from a baseline period prior to lickspout placement were observed in the 1 min period following lickspout placement, and in the 1 min period following access to water. Example spike waveforms (averaged across 10 spikes) confirm that endogenous spike waveform shapes (black) remained stable across the recording and match the shape of photostimulation-evoked spike waveforms (blue).

(D) Time course of percent change in firing from pre-lickspout baseline (green vertical line) for all identified VPpp neurons recorded during this task (n = 14). Short vertical black lines denote the onset of water availability. Example neurons in (B) and (C) are labeled “#1” and “#2. ” The dashed gray line indicates the end of the period of manual lickspout positioning and adjustment, a period excluded from subsequent analyses.

(E) Percent change in spiking rate of VPpp neurons and unidentified SON neurons for different drinking periods. Error bars denote SEM. Analyses included n = 10 VPpp neurons and n = 27 unidentified SON neurons for the pre-drinking period (see Experimental Procedures). For all other periods, all VPpp neurons (n = 14) and unidentified neurons (n = 34) were included. Paired t tests: *p < 0.05, **p < 0.01.

See also Figures S1–S3.

Spiking Responses of Identified SON VPpp Neurons in Awake Mice

We then recorded spiking activity in individual SON VPpp neurons using optetrodes (see Mandelblat-Cerf et al., 2015, and Experimental Procedures). To selectively express ChR2 in SON-VPpp cell bodies, we injected AAV9-FLEX-hSYN-ChR2-mCherry into the SON (Figure 1A). We recorded spiking activity in 48 neurons (21 sessions in 4 mice). Of these neurons, 14 were identified as VPpp neurons based on a significant increase in firing in response to optogenetic photostimulation (Figure S2). All optetrode recordings were performed in water-restricted mice habituated to head restraint and automated water delivery via a lickspout.

Spiking activity was recorded (1) prior to presentation of any water-predicting cues (“baseline period”), (2) following positioning of the lickspout in front of the snout such that licks did not result in water delivery, and (3) during a period in which the mouse licked to receive water. The timing of lickspout placement and of subsequent water availability was varied across sessions to disambiguate pre-drinking VPpp neuron responses to the water-predicting cue (lickspout placement) from subsequent responses to drinking onset. Example recordings from two VPpp neurons are shown in Figures 1B and 1C (see also Figures S2A and S2B). Baseline firing rates in VPpp neurons in water-restricted mice were quite high (range: 4–60 spikes/s, mean ± SD: 22 ± 14 spikes/s, see Figure S2C). Both example VPpp neurons showed a dramatic decrease in firing rate within seconds of lickspout placement, and again following onset of water availability (Figures 1B and 1C).

Inspection of time courses of changes in spiking activity relative to baseline (Figure 1D; green vertical line) revealed a decrease in firing in all 14 VPpp neurons (blue, decrease in firing; white, no change; red, increase). We quantified changes from baseline firing in six windows of time (Figure 1E): (1) between the end of lickspout positioning and the onset of drinking, a period in which mice lick but no water is delivered, and (2–6) at 0–0.5, 0.5–1, 1–3, 3–5, and 5–7 min after drinking onset—periods in which mice licked for water in a sustained manner. We found a significant (~25%) average drop in spiking following lickspout presentation, prior to drinking onset (Figure 1E; paired t test: p = 0.002). We also observed an additional ~10% decrease in average VPpp neuron firing in the 30 s following drinking onset (paired t test relative to baseline: p < 0.0001), and an additional drop of 15% within the next 150 s (paired t test relative to baseline: p < 0.001; relative to pre-drinking period: p = 0.01). Interestingly, average VPpp neuron firing then remained at the same low level throughout subsequent periods of time. In contrast to the robust decreases in firing observed across VPpp neurons, no significant changes in average firing were observed across nearby, unidentified neurons (Figures 1E, S2C, and S2D), consistent with diverse functions attributed to non-VPpp neurons in the SON.

Responses across individual VPpp neuron responses were relatively homogeneous. Of the 10/14 identified VPpp neurons for which sufficient data existed between termination of lickspout placement and onset of drinking, 90% showed a significant water cue-evoked drop in mean firing prior to any water ingestion (Figure S2D, K-S tests, p < 0.01 after Bonferroni correction). In contrast, only 47% (14/30) of unidentified SON neurons showed a significant drop in mean firing during this period. Similar results were observed at later time periods following drinking onset. In addition, similar drops in activity following water cues and water ingestion were observed in additional optetrode recordings from five confirmed PVH-VPpp neurons (Figures S2E–S2H).

In separate experiments, we found that the elevation in plasma osmolality following water restriction only began to return to baseline levels after more than 10 min of water ingestion (Figure S3). Thus, both cue-evoked and early drinking-evoked decreases in VPpp neuron firing were “presystemic” in nature, as they occurred prior to any drinking-induced changes in plasma osmolality and, consequently, prior to any systemic feedback signals.

Decreases in SON VPpp Neuron Activity across Hyperosmotic Contexts

To assess the generality of these presystemic changes in VPpp neuron activity across hyperosmolar contexts and across ingestive behaviors, we used fiber photometry to monitor the aggregate calcium activity of populations of VPpp neurons in freely moving mice (Figure 2A; see Garfield et al., 2016, and Experimental Procedures). Experiments were conducted in the home-cage, and mice were habituated to drinking from a water bowl that was placed in the cage midway through each session.

Figure 2. Water Cues and Drinking Both Cause Rapid Drops in VPpp Population Activity across Multiple Hyperosmolar Contexts.

Figure 2

Open in a new tab

(A) Upper panel: GCaMP6 fiber photometry measurements of population activity in VPpp neurons via a 200 μm fiber in SON. Lower panel: example VPpp population activity after 24 hr of water deprivation, demonstrating a fast drop within seconds of presentation of a water bowl (green vertical line) and an additional drop at drinking onset (gray vertical line). Inset: zoom-in of activity surrounding bowl placement and drinking onset.

(B) Change in VPpp neuron activity across drinking periods, demonstrating a significant drop before drinking onset (n = 5 mice).

(C) In contrast to water-restricted mice (blue trace), sated mice (gray trace, n = 3) with free access to water showed no decrease in VPpp neuron activity following water bowl presentation (p > 0.15 for periods from 0–30 s and 30–60 s).

(D) As with water restriction, water cues and drinking following consumption of dry food both caused a rapid drop in VPpp population activity (n = 5 mice).

(E) In contrast to these sustained responses to drinking (light blue trace), unexpected presentation of an empty bowl lacking water (gray trace) caused a transient drop in VPpp neuron activity (at both 5–10 and 10–15 s after placement; *paired t test, p < 0.05, Bonferroni corrected for multiple comparisons) that promptly returned to baseline levels.

Paired t tests in (B) and (D): *p < 0.05, **p < 0.01. Bonferroni corrected for multiple comparisons. Error bars denote SEM. See also Figure S3.

Photometry recordings in freely moving mice following 24 hr of water deprivation revealed a significant drop in population VPpp activity following presentation of the water bowl, prior to drinking (paired t test relative to baseline: p = 0.049), a further drop within 30 s of drinking onset (paired t test relative to baseline: p = 0.0014; relative to pre-drinking period: p = 0.003), and a maximal reduction in activity reached within ~7 min of drinking onset (Figures 2A and 2B). While plasma osmolality—the main systemic feedback signal—decreased from 10–15 min post drinking onset (Figure S3A), no additional significant decrease in VPpp neuron activity was observed during this later period. Finally, the observed drops in VPpp neuron activity were state dependent, as the same presentation of water did not cause a significant decrease in VPpp neuron activity in mice with ad libitum access to water prior to the recording session (Figure 2C, gray trace).

To test the generality of rapid, drinking-induced decreases in VPpp neuron activity across other hyperosmolar contexts, we next examined responses of VPpp neurons to water ingestion following consumption of dry food. Sustained consumption of dry food without access to water elevated plasma osmolality in mice (Figure S3C, “post-feeding” period; see also Zimmerman et al., 2016). Further, consumption of 1 g of chow appeared to increase thirst, as reflected by a robust increase in the motivational drive to work to obtain water (Figure S3D). Accordingly, we found water-predicting cues and drinking following consumption of 1 g of chow resulted in rapid drops in activity in VPpp neurons (Figure 2D), similar to those observed following water restriction (Figure 2B). Average VPpp neuron activity decreased significantly in the period following presentation of the water bowl but prior to drinking onset (paired t test: p = 0.018) and continued to drop in the 30 s following ingestion of water (paired t test versus baseline: p = 0.0002; relative to pre-drinking period: p = 0.001). In subsequent sessions, presentation of the same bowl lacking water resulted in a significant, but transient, decrease in VPpp neuron activity, followed by a return to baseline activity (Figure 2E, black trace; a significant drop occurred from 0–15 s following bowl placement compared to baseline, p = 0.011).

Increases in SON VPpp Neuron Activity during Dry Food Consumption

We wondered whether similarly rapid predictive increases in VPpp neuron activity might occur prior to and during consumption of dry food. While presystemic increases in plasma vasopressin release have been demonstrated in rats in response to a hyperosmotic load (gastric infusion of NaCl, Choi-Kwon et al., 1990; Stricker and Hoffmann, 2007; Stricker et al., 2002), the dynamics of presystemic changes in VPpp neuron activity in natural feeding contexts remain largely unknown (Stricker and Hoffmann, 2007). Thus, we performed fiber photometry recordings of the activity of VPpp neurons in food-restricted mice during (1) presentation of a bowl associated (across training sessions) with the presence of food, and subsequent ad libitum access to chow for at least 15 min (Figure 3A), (2) presentation of a large chow food pellet (250 mg, Figure 3B; consumed within 90–180 s), and (3) presentation of a small chow food pellet (15 mg, consumed within seconds) or a non-food item of similar size and weight (Figure 3C).

Figure 3. Feeding Onset and Offset Also Induce Rapid VPpp Population Responses.

Figure 3

Open in a new tab

(A–C) VPpp population response to ad libitum ingestion of dry food from a food bowl (A), and to presentation of 250 mg food pellets (B), 15 mg pellets (C, black) and intermingled presentation of non-food items of similar size (C, red). Upper traces: examples. Lower traces: average responses across mice. Shaded error bars denote SEM (n = 5 mice). Beige shaded regions denote periods of food availability.

(D–F) Average VPpp population responses to feeding. Error bars denote SEM. *p < 0.05, **p < 0.01, paired t test with Bonferroni correction for multiple comparisons. Lower asterisks indicate significant differences from baseline activity. Upper asterisks indicate significant differences from activity in the period prior to feeding offset. n.s.: not significant. See also Figure S3.

We compared average responses across mice to food cues and to food ingestion (Figures 3D–3F), similar to analyses of responses to water cues and to water ingestion (Figures 2B and 2D). The population activity of VPpp neurons rapidly increased within 30 s of onset of ad libitum chow feeding (Figure 3D, paired t test: p = 0.014). This increase occurred prior to any significant rise in plasma osmolality, which began more than 2 min after feeding onset (Figure S3E). To more precisely estimate the temporal dynamics of changes in VPpp neuron activity prior to and during feeding, we measured responses to presentation of individual food pellets. VPpp neuron activity increased within 10 s of onset of consumption of either a large chow pellet (Figures 3B and 3E; p = 0.018) or of a small chow pellet (Figures 3C and 3F; p = 0.008). In contrast, non-food items did not cause any change in activity (Figures 3C and 3F, red traces).

These recordings also revealed sharp drops in VPpp neuron activity at the termination of consumption of individual pellets. Following consumption of a small chow pellet, VPpp neuron activity returned to baseline within 20–30 s (Figures 3C and 3F). Similarly, within 30 s of termination of consumption of a large chow pellet, we observed a resetting of VPpp neuron activity to an intermediate level that was higher than baseline activity but lower than during feeding (Figures 3B and 3E). In contrast, VPpp neuron activity remained elevated following ad libitum access to chow for at least 15 min, even after the food bowl was removed (Figures 3A and 3D; comparison of 2 min periods pre- versus post-bowl removal).

Robust Pre-ingestive VPpp Population Responses to Water Cues but Not to Food Cues

Strikingly, in contrast to the pre-ingestive change in VPpp neuron activity following presentation of a bowl previously associated with availability of water, there was no significant pre-ingestive change in activity following the presentation of a bowl previously associated with availability of food (Figures 3A and 3D, p > 0.05). To further assess this asymmetry in pre-ingestive responses to food versus water, we assessed VPpp neuron activity in individual trials across all mice (Figures 4A–4E). Trials were aligned to time of item placement (left column) or to consumption onset (right column) and sorted according to the delay between item placement and a conservative estimate of ingestion onset (black tick marks indicate estimated time of first contact with food or water). Increases in activity following presentation of a food-associated bowl (Figure 4A), a large pellet (Figure 4B), or a small pellet (Figure 4C) were reliable across single trials. However, prior to feeding onset, no increase was evident in any trials. In contrast, prior to drinking onset, placement of a water-associated bowl (post 24 hr water deprivation, Figure 4D, or post feeding, Figure 4E) resulted in highly reliable drops in activity across individual trials (see also Figure 4F).

Figure 4. Highly Reliable Pre-ingestive VPpp Population Responses to Water Cues but Not to Food Cues.

Figure 4

Open in a new tab

(A–E) Single-trial time courses of increases (red), decreases (blue), or no change (white) in VPpp population activity relative to the period prior to food item placement (A–C) or water bowl placement (D and E). Trials are sorted according to latency from item placement to estimated ingestion onset (vertical black ticks). Left column is aligned to item presentation (green vertical line); right column is aligned to ingestion onset. (A–C) Ingestion of food from a bowl (A), a large food pellet (B), or a small food pellet (C) following food restriction. (D and E) Ingestion of water following 24 hr water deprivation (D) or following consumption of dry food (E). Activity increased within 20 s of consumption for 100% (12/12) of food bowl events, 99% (116/117) of large pellet events, and for 93% (135/145) of small pellet events. In contrast, activity decreased in 96% (51/53) of feeding-induced drinking events and 100% (12/12) of water deprivation-induced drinking events. We then considered changes in activity following item presentation but prior to ingestion (for all events in which this pre-consumption window lasted at least 5 s). The water cue evoked a decrease in activity in 89% (32/36) and 91% (10/11) of trials that followed chow feeding and water deprivation, respectively. In contrast, presentation of food cues did not evoke consistent increases in activity in food-restricted mice prior to ingestion onset (51%, 35/68; 51%, 27/53; and 33%, 4/12).

(F1) Similarly, average cue-evoked changes in activity across trials revealed a significant drop following water bowl presentation but prior to drinking (p < 0.01), but no response to presentation of food items prior to feeding (all p values > 0.1). Error bars denote SEM.

(F2) In contrast, significant changes from pre-cue baseline to the 10 s period following onset of ingestion were observed for both feeding and drinking (**p < 0.01). Error bars denote SEM.

(G) Schematic illustrating the various presystemic inputs that are likely to contribute to the fast increases and decreases in VPpp neuron activity during ingestive behavior. “X” indicates that pre-ingestive food cue signals to do not appear to affect VPpp neuron activity.

Finally, we further investigated the nature of the observed pre-ingestive response to water availability by using an automated classical conditioning paradigm (Figure S4). Freely moving water-restricted mice (n = 3) instrumented for fiber photometry recordings were exposed to a 5 s stimulus (a bright light), prior to and following association of this conditioned stimulus (CS) with water delivery from a lickspout. Prior to the learned association, VPpp neurons did not exhibit a significant response to CS presentation (paired t test: p > 0.1 for all mice). However, following association of the CS with water delivery, VPpp neurons showed a significant pre-ingestive drop in activity in all three mice (paired t tests: p = 0.043, 0.001, and 0.001).

DISCUSSION

Here, we monitored the real-time activity of genetically defined, pituitary-projecting neuroendocrine motor neurons that release vasopressin to control renal water excretion (Bourque, 2008; Nielsen et al., 1995; Sands et al., 2011; Watts, 2015). Using electrophysiological recordings in genetically identified SON pituitary-projecting vasopressin (VPpp) neurons in water-restricted mice, we observed rapid decreases in spiking within seconds of presentation of cues signaling water availability, beginning prior to water ingestion. We confirmed this finding in freely moving, water-restricted mice by using fiber photometry across several hyperosmotic contexts. In contrast, ingestion of dry food—a hyperosmotic challenge—elicited rapid increases in VPpp neuron activity prior to any increase in plasma osmolality. Surprisingly, in contrast to the water cue-induced drops in activity prior to drinking, we did not observe food cue-induced increases in activity prior to feeding onset (for summary, see Figure 4G).

Previous studies showed that plasma vasopressin levels drop within minutes of water ingestion, prior to systemic changes in plasma osmolality (Geelen et al., 1984; Huang et al., 2000; Stricker and Hoffmann, 2007), yet the fine timing of changes in SON neuron activity (Arnauld and du Pont, 1982) in identified VPpp neurons remained unclear. We found that individual identified VPpp neurons showed consistent, rapid drops in firing within seconds of water availability and ingestion. These drops were highly reliable, as evident in population responses across individual trials, sessions, and mice. Interestingly, we found that VPpp neuron spiking quickly reached a lower, albeit non-zero, steady-state level during water ingestion and did not appear to change further at later times at which these neurons are expected to receive systemic feedback regarding drops in plasma osmolality. Thus, VPpp neurons appear to accurately predict the consequences of drinking to satiety and quickly reset their activity to a level that persists following satiety.

Our data support the hypothesis that multiple neuronal inputs relay distinct cue- and ingestion-related information to VPpp neurons regarding the availability of water and food, leading to rapid, robust changes in VPpp neuron activity. These rapid changes in VPpp neuron activity are likely to affect water retention by the kidney within tens of seconds, for the following reasons: (1) vasopressin has a relatively short half-life in blood plasma (Lauson, 1967), (2) blood circulates throughout the mouse body within 15 s (Debbage et al., 1998), and (3) kidney collecting duct epithelial cells react to changes in external vasopressin concentration within 10–20 s (Chou et al., 2000). Thus, at the onset of ingestive behaviors, neuroendocrine VPpp neurons may respond proactively to the expected upcoming systemic changes in plasma osmolality in order to dampen transient osmotic stress.

Comparison to Predictive Signals in the Control of Energy Balance

Other genetically defined populations of homeostasis-promoting neurons in the hypothalamus—agouti-related peptide (AgRP) and pro-opiomelanocortin (POMC) neurons—have been shown to exhibit rapid, feeding-related changes in activity using similar methods for monitoring spiking and calcium activity. AgRP neurons increase their activity during caloric restriction and show rapid drops in activity both prior to and during food ingestion (Chen et al., 2015; Betley et al., 2015; Mandelblat-Cerf et al., 2015; Burnett et al., 2016). This rapid drop in AgRP neuron activity upon food presentation may play a role in preventing overconsumption in the period prior to systemic feedback (Chen et al., 2015; Mandelblat-Cerf et al., 2015). Activation of AgRP neurons was reported to carry a negative valence when food is not available (Betley et al., 2015) and to promote appetitive behavior when food subsequently becomes available (Chen et al., 2016). Thus, the rapid drop in AgRP neuron activity with food availability may act as a teaching signal to the many target brain areas that receive dense AgRP neuron projections (Betley et al., 2013). In contrast, VPpp neurons do not send substantial projections to other brain areas (Figure S1) and have not been shown to acutely influence behavior, suggesting that the fast changes in VPpp neuron activity primarily act to mitigate under-shoots or overshoots in plasma osmolality during ingestive behavior.

Bidirectional yet Asymmetric Responses to Drinking and Feeding

VPpp neurons also showed a rapid increase in activity at the onset of dry food ingestion, suggesting that the same neuroendocrine neurons are capable of rapid, bidirectional modulation of kidney function, beginning in the period prior to any systemic feedback. Interestingly, in contrast to the short latency neural response to water cues prior to water ingestion, we did not observe any response to food cues prior to food ingestion. We speculate that water ingestion may have more predictable effects on plasma osmolality, while food ingestion may have more variable effects depending on the nature of the food. In this way, it may be more difficult to anticipate expected osmotic load prior to food consumption. Note that while our experiments involving separate access to water and dry food are ecologically relevant, field mice can also extract water from many sources of food (http://www.in.gov/isdh/23256.htm)—a situation that likely requires a more nuanced estimation of predicted osmotic load.

Potential Contributors to Presystemic Signals in VPpp Neurons

The fast presystemic changes in VPpp neuron activity that we observed likely arise from many sources (Figure 4G). Drinking-related oropharyngeal signals such as swallowing were shown to presystemically inhibit VP neurons in many species (Arnauld and du Pont, 1982; Geelen et al., 1984; Thrasher et al., 1987). In rodents, however, only ingestion of water, but not saline, induced rapid decreases in vasopressin (Huang et al., 2000), suggesting that other mechanisms are also involved. Presystemic signals may stem from taste and thermosensation (Zimmerman et al., 2016). Additionally, osmosensitive vagal afferents that innervate the liver can sense early changes in hepatic circulation and quickly relay signals to the central nervous system (Lechner et al., 2011). Similarly, signaling by intestinal vagal afferents in response to hypo- and hypertonic fluids are known to influence vasopressin release (Baertschi and Pence, 1995; Stricker et al., 2002). Notably, thirst-promoting neurons in the subfornical organ (SFO), which are known to project to the SON (Gash and Boer, 1987), were recently shown to exhibit fast increases and decreases in activity in response to feeding and drinking, respectively (Zimmerman et al., 2016). However, in contrast to VPpp neurons, these SFO neurons did not respond to non-orofacial water-predicting cues, further supporting the existence of functionally diverse sources of pre-systemic input onto VPpp neurons.

As with drinking responses, the rapid VPpp responses at onset of dry food consumption are also likely driven, in part, by oropharyngeal and taste signals. The dynamics of VPpp neuron activity during and at the offset of feeding also provide clues as to the drivers of this presystemic activity. After consuming a small food pellet (Figure 3C), VPpp neuron activity promptly returned to baseline within seconds, further suggesting that the driver of the fast feeding-related increase in VPpp activity may be oropharyngeal in nature. In contrast, VPpp neuron activity only partially returned to baseline following termination of a minutes-long bout of consumption of a larger pellet, most likely due to more distal presystemic signals and to early systemic signals.

We have demonstrated bidirectional presystemic changes in the activity of identified VPpp neurons at a timescale of seconds following the onset of feeding and drinking, as well as in response to water-predicting sensory cues. In future, this genetically accessible entry point in the osmoregulatory circuit can be mined to determine the upstream pre-motor circuits responsible for these rapid and opposing effects on VPpp neuron activity (Bourque, 2008; Stricker and Hoffmann, 2007), as well as potential hypothalamic areas (Powley, 1977), limbic areas (Greenwood et al., 1991; Roozendaal et al., 1990), and cortical areas (Levinthal and Strick, 2012) involved in learning and top-down control of neuroendocrine motor outflow. Thus, our findings set the stage for understanding the neglected but important circuitry regulating rapid motor control of the kidney and its possible role in minimizing osmotic stress during self-initiated ingestive behaviors.

EXPERIMENTAL PROCEDURES

All animal care and experimental procedures were approved by the Beth Israel Deaconess Medical Center Institutional Animal Care and Use Committee. For details on all methods, please see Supplemental Experimental Procedures.

Supplementary Material

1
2

Highlights.

  • Recordings from vasopressin neuroendocrine motor neurons (VPpp) in behaving mice

  • Feeding, but not cues predicting food, increase VPpp neuron activity within seconds

  • Drinking and cues predicting water reduce VPpp neuron activity within seconds

  • Drinking-related reductions in activity reach steady state prior to systemic feedback

Acknowledgments

We would like to thank Dr. A. Sugden for assistance with breakpoint analysis experiments and members of the M.L.A. and B.B.L. labs and Drs. C. Saper, J. Majzoub, S. Liberles, J. Geerling, and Y. Livneh for helpful discussion. We thank Drs. Jayaraman, Kerr, Kim, Looger, and Svoboda and the GENIE Project at JFRC (HHMI) for distribution of GCaMP6s. Support was provided by a Charles A. King Trust Postdoctoral Fellowship (Y.M.-C.), a Davis Family Foundation Postdoctoral Fellowship (C.R.B.), NIH R01 DK075632, R01 DK096010, R01 DK089044, P30 DK046200, and P30 DK057521 (B.B.L.), NIH R01 DK109930, DP2 DK105570, the Pew Scholars Program in the Biomedical Sciences, the Klarman Family Foundation, and the Smith Family Foundation (M.L.A.).

Footnotes

AUTHOR CONTRIBUTIONS

Y.M.-C., B.B.L., and M.L.A. designed the experiments and wrote the manuscript. Y.M.-C. collected and analyzed electrophysiology and photometry data. A.K. conducted experiments involving blood osmolality and viral tracing. C.R.B. constructed the photometry setup. S.S. designed the progressive ratio task and acquired associated data. B.A.T. provided the rabies virus.

References

  1. Armstrong WE. Morphological and electrophysiological classification of hypothalamic supraoptic neurons. Prog Neurobiol. 1995;47:291–339. [PubMed] [Google Scholar]
  2. Arnauld E, du Pont J. Vasopressin release and firing of supraoptic neurosecretory neurones during drinking in the dehydrated monkey. Pflugers Arch. 1982;394:195–201. doi: 10.1007/BF00589091. [DOI] [PubMed] [Google Scholar]
  3. Baertschi AJ, Pence AR. Gut-brain signaling of water absorption inhibits vasopressin in rats. Am J Physiol. 1995;268:R236–247. doi: 10.1152/ajpregu.1995.268.1.R236. [DOI] [PubMed] [Google Scholar]
  4. Betley JN, Cao ZFH, Ritola KD, Sternson SM. Parallel, redundant circuit organization for homeostatic control of feeding behavior. Cell. 2013;155:1337–1350. doi: 10.1016/j.cell.2013.11.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Betley JN, Xu S, Cao ZFH, Gong R, Magnus CJ, Yu Y, Sternson SM. Neurons for hunger and thirst transmit a negative-valence teaching signal. Nature. 2015;521:180–185. doi: 10.1038/nature14416. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bourque CW. Central mechanisms of osmosensation and systemic osmoregulation. Nat Rev Neurosci. 2008;9:519–531. doi: 10.1038/nrn2400. [DOI] [PubMed] [Google Scholar]
  7. Burnett CJ, Li C, Webber E, Tsaousidou E, Xue SY, Bruning JC, Krashes MJ. Hunger-driven motivational state competition. Neuron. 2016;92:187–201. doi: 10.1016/j.neuron.2016.08.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chen Y, Lin YC, Kuo TW, Knight ZA. Sensory detection of food rapidly modulates arcuate feeding circuits. Cell. 2015;160:829–841. doi: 10.1016/j.cell.2015.01.033. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen Y, Lin YC, Zimmerman CA, Essner RA, Knight ZA. Hunger neurons drive feeding through a sustained, positive reinforcement signal. eLife. 2016;5:1–21. doi: 10.7554/eLife.18640. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Choi-Kwon S, McCarty R, Baertschi AJ. Splanchnic control of vasopressin secretion in conscious rats. Am J Physiol. 1990;259:E19–E26. doi: 10.1152/ajpendo.1990.259.1.E19. [DOI] [PubMed] [Google Scholar]
  11. Chou CL, Yip KP, Michea L, Kador K, Ferraris JD, Wade JB, Knepper MA. Regulation of aquaporin-2 trafficking by vasopressin in the renal collecting duct. Roles of ryanodine-sensitive Ca2+ stores and calmodulin. J Biol Chem. 2000;275:36839–36846. doi: 10.1074/jbc.M005552200. [DOI] [PubMed] [Google Scholar]
  12. Cohen JY, Haesler S, Vong L, Lowell BB, Uchida N. Neuron-type-specific signals for reward and punishment in the ventral tegmental area. Nature. 2012;482:85–88. doi: 10.1038/nature10754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Debbage PL, Griebel J, Ried M, Gneiting T, DeVries A, Hutzler P. Lectin intravital perfusion studies in tumor-bearing mice: micrometer-resolution, wide-area mapping of microvascular labeling, distinguishing efficiently and inefficiently perfused microregions in the tumor. J Histochem Cytochem. 1998;46:627–639. doi: 10.1177/002215549804600508. [DOI] [PubMed] [Google Scholar]
  14. Garfield AS, Shah BP, Burgess CR, Li MM, Li C, Steger JS, Madara JC, Campbell JN, Kroeger D, Scammell TE, et al. Dynamic GABAergic afferent modulation of AgRP neurons. Nat Neurosci. 2016 doi: 10.1038/nn.4392. http://dx.doi.org/10.1038/nn.4392. [DOI] [PMC free article] [PubMed]
  15. Gash D-M, Boer GJ. Vasopressin: Principles and Properties. Springer; 1987. [Google Scholar]
  16. Geelen G, Keil LC, Kravik SE, Wade CE, Thrasher TN, Barnes PR, Pyka G, Nesvig C, Greenleaf JE. Inhibition of plasma vasopressin after drinking in dehydrated humans. Am J Physiol. 1984;247:R968–R971. doi: 10.1152/ajpregu.1984.247.6.R968. [DOI] [PubMed] [Google Scholar]
  17. Greenwood RS, Meeker RB, Hayward JN. Amygdala kindling elevates plasma vasopressin. Brain Res. 1991;538:9–14. doi: 10.1016/0006-8993(91)90369-7. [DOI] [PubMed] [Google Scholar]
  18. Harris GW. Neural control of the pituitary gland. I The neurohypophysis. BMJ. 1951;2:559–564. doi: 10.1136/bmj.2.4731.559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hatton GI. Emerging concepts of structure-function dynamics in adult brain: the hypothalamo-neurohypophysial system. Prog Neurobiol. 1990;34:437–504. doi: 10.1016/0301-0082(90)90017-b. [DOI] [PubMed] [Google Scholar]
  20. Huang W, Sved AF, Stricker EM. Water ingestion provides an early signal inhibiting osmotically stimulated vasopressin secretion in rats. Am J Physiol Regul Integr Comp Physiol. 2000;279:R756–R760. doi: 10.1152/ajpregu.2000.279.3.R756. [DOI] [PubMed] [Google Scholar]
  21. Lauson HD. Metabolism of antidiuretic hormones. Am J Med. 1967;42:713–744. doi: 10.1016/0002-9343(67)90091-5. [DOI] [PubMed] [Google Scholar]
  22. Lechner SG, Markworth S, Poole K, Smith ESJ, Lapatsina L, Frahm S, May M, Pischke S, Suzuki M, Ibañez-Tallon I, et al. The molecular and cellular identity of peripheral osmoreceptors. Neuron. 2011;69:332–344. doi: 10.1016/j.neuron.2010.12.028. [DOI] [PubMed] [Google Scholar]
  23. Leng G, Brown CH, Russell JA. Physiological pathways regulating the activity of magnocellular neurosecretory cells. Prog Neurobiol. 1999;57:625–655. doi: 10.1016/s0301-0082(98)00072-0. [DOI] [PubMed] [Google Scholar]
  24. Levinthal DJ, Strick PL. The motor cortex communicates with the kidney. J Neurosci. 2012;32:6726–6731. doi: 10.1523/JNEUROSCI.0406-12.2012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Mandelblat-Cerf Y, Ramesh RN, Burgess CR, Patella P, Yang Z, Lowell BB, Andermann ML. Arcuate hypothalamic AgRP and putative POMC neurons show opposite changes in spiking across multiple timescales. eLife. 2015;4:e07122. doi: 10.7554/eLife.07122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Nielsen S, Chou CL, Marples D, Christensen EI, Kishore BK, Knepper MA. Vasopressin increases water permeability of kidney collecting duct by inducing translocation of aquaporin-CD water channels to plasma membrane. Proc Natl Acad Sci USA. 1995;92:1013–1017. doi: 10.1073/pnas.92.4.1013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Pavlov J. The Work of Digestive Glands. Charles Griffin; 1902. [Google Scholar]
  28. Pei H, Sutton AK, Burnett KH, Fuller PM, Olson DP. AVP neurons in the paraventricular nucleus of the hypothalamus regulate feeding. Mol Metab. 2014;3:209–215. doi: 10.1016/j.molmet.2013.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Poulain DA, Wakerley JB. Electrophysiology of hypothalamic magnocellular neurones secreting oxytocin and vasopressin. Neuroscience. 1982;7:773–808. doi: 10.1016/0306-4522(82)90044-6. [DOI] [PubMed] [Google Scholar]
  30. Powley TL. The ventromedial hypothalamic syndrome, satiety, and a cephalic phase hypothesis. Psychol Rev. 1977;84:89–126. [PubMed] [Google Scholar]
  31. Roozendaal B, Oldenburger WP, Strubbe JH, Koolhaas JM, Bohus B. The central amygdala is involved in the conditioned but not in the meal-induced cephalic insulin response in the rat. Neurosci Lett. 1990;116:210–215. doi: 10.1016/0304-3940(90)90412-3. [DOI] [PubMed] [Google Scholar]
  32. Sands JM, Blount MA, Klein JD. Regulation of renal urea transport by vasopressin. Trans Am Clin Climatol Assoc. 2011;122:82–92. [PMC free article] [PubMed] [Google Scholar]
  33. Stricker EM, Hoffmann ML. Presystemic signals in the control of thirst, salt appetite, and vasopressin secretion. Physiol Behav. 2007;91:404–412. doi: 10.1016/j.physbeh.2007.04.007. [DOI] [PubMed] [Google Scholar]
  34. Stricker EM, Callahan JB, Huang W, Sved AF. Early osmo-regulatory stimulation of neurohypophyseal hormone secretion and thirst after gastric NaCl loads. Am J Physiol Regul Integr Comp Physiol. 2002;282:R1710–R1717. doi: 10.1152/ajpregu.00548.2001. [DOI] [PubMed] [Google Scholar]
  35. Thrasher TN, Keil LC, Ramsay DJ. Drinking, oropharyngeal signals, and inhibition of vasopressin secretion in dogs. Am J Physiol. 1987;253:R509–R515. doi: 10.1152/ajpregu.1987.253.3.R509. [DOI] [PubMed] [Google Scholar]
  36. Watts AG. 60 YEARS OF NEUROENDOCRINOLOGY: The structure of the neuroendocrine hypothalamus: the neuroanatomical legacy of Geoffrey Harris. J Endocrinol. 2015;226:T25–T39. doi: 10.1530/JOE-15-0157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Woods SC. The eating paradox: how we tolerate food. Psychol Rev. 1991;98:488–505. doi: 10.1037/0033-295x.98.4.488. [DOI] [PubMed] [Google Scholar]
  38. Zimmerman CA, Lin YC, Leib DE, Guo L, Huey EL, Daly GE, Chen Y, Knight ZA. Thirst neurons anticipate the homeostatic consequences of eating and drinking. Nature. 2016;537:680–684. doi: 10.1038/nature18950. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS836099-supplement-1.pdf (11.6MB, pdf)
2
NIHMS836099-supplement-2.pdf (13.2MB, pdf)

RESOURCES