Abstract
The regulation of food intake and energy balance relies on the dynamic integration of exteroceptive and interoceptive signals monitoring nutritional, metabolic, cognitive and emotional states. The paraventricular thalamus (PVT) is a central hub that, by integrating sensory, metabolic and emotional states, may contribute to the regulation of feeding and homeostatic/allostatic processes. However, the underlying PVT circuits remain still elusive. Here, we aimed at unraveling the role of catecholaminergic (CA) inputs to the PVT in scaling feeding and metabolic efficiency. First, using region-specific retrograde disruption of CA projections, we show that PVT CA inputs mainly arise from the hindbrain, notably the locus coeruleus (LC) and the nucleus tractus solitarius (NTS). Second, taking advantage of integrative calorimetric measurements of metabolic efficiency, we reveal that CA inputs to the PVT scale adaptive feeding and metabolic responses in environmental, behavioral, physiological and metabolic stress-like contexts. Third, we show that hindbrainTH→PVT inputs contribute in modulating the activity of PVT as well as lateral (LH) and dorsomedial (DMH) hypothalamic neurons.
In conclusion, this study, by assessing the key role of CA inputs to the PVT in scaling homeostatic/allostatic regulations of feeding patterns, reveals the integrative and converging hindbrainTH→PVT paths that contribute to whole-body metabolic adaptations in stress-like contexts.
Key points
The paraventricular thalamus (PVT) is known to receive projections from the hindbrain. Here, we confirm and further extend current knowledge on the existence of hindbrainTH→PVT catecholaminergic (CA) inputs, notably from the locus coeruleus (LC) and the nucleus tractus solitarius (NTS), with the NTS representing the main source.
Disruption of hindbrainTH→PVT inputs contribute to the modulation of PVT-neurons activity.
HindbrainTH→PVT inputs scale feeding strategies in environmental, behavioral, physiological and metabolic stress-like contexts.
HindbrainTH→PVT inputs participate in regulating metabolic efficiency and nutrient partitioning in stress-like contexts.
HindbrainTH→PVT, directly and/or indirectly, contribute in modulating the downstream activity of lateral (LH) and dorsomedial (DMH) hypothalamic neurons.
Introduction
In mammals, the regulation of food intake intricately relies on the orchestration of several signals mirroring the dynamic integration of interoceptive and exteroceptive (environment) states (Sweeney & Yang, 2017). Indeed, emotional states (stress, anxiety, motivation), by modulating the activity of central neuronal hubs, thoroughly scale the regulation of food intake and the establishment of feeding habits (Ulrich-Lai et al., 2015). This is particularly evident in stress-related contexts where emotional states compete with the homeostatic regulation of feeding (Maniam & Morris, 2012; Herzog, 2020), thereby leading to feed-forward allostatic adaptations (stability through changes) which may culminate in psychiatric and metabolic disorders. These observations support the idea that emotional and homeostatic states may share similar, although not identical, neuronal networks (Sweeney & Yang, 2017). Nonetheless, the systems underlying the functional connection between these states remain poorly understood.
Emerging evidence strongly suggests that the paraventricular nucleus of the thalamus (PVT), a dorsal midline thalamic relay, may represent a functional hot-spot where interoceptive and exteroceptive stimuli may converge to orchestrate the selection of adaptive and appropriate behavioral responses aimed at regulating homeostatic and cognitive processes (Penzo & Gao, 2021). Given their elaborated connectivity (Kirouac, 2015), PVT excitatory (glutamate) neurons are well positioned to serve as functional integrators of orexigenic and anorexigenic stimuli (Ong et al., 2017; Meffre et al., 2019), glucose fluctuations (Labouèbe et al., 2016; Kessler et al., 2021), learning and memory processes (Do-Monte et al., 2015; Penzo et al., 2015) as well as emotional states (Heydendael et al., 2011; Barson et al., 2020; Pliota et al., 2020). This plethora of PVT-related brain functions highly relies on different afferent projections which, by carrying distinct neurochemical information, synergistically scale the activity of PVT-neurons and their downstream connected regions. Among the different afferents, the PVT also harbors a dense plexus of catecholaminergic (CA) fibers mainly arising from the hindbrain (Schroeter et al., 2000; Beas et al., 2018; Sofia Beas et al., 2020; Wang et al., 2021b) and only few scattered CA fibers from the hypothalamus (Li et al., 2014). In addition, recent reports have indicated that brain CA (norepinephrine, NE, and/or dopamine, DA), by modulating different homeostatic functions (i.e. wakefulness, arousal, glucoprivation-induced food seeking), may represent key neuromodulators of the PVT (Beas et al., 2018; Sofia Beas et al., 2020; Wang et al., 2021b). Moreover, brain CA are also important contributors to the regulation of stress-like responses (Valentino & Van Bockstaele, 2008; Kvetnansky et al., 2009) which ultimately impact, directly and/or indirectly, on feeding patterns and behaviors (Xu et al., 2019; Qu et al., 2020).
Indeed, the PVT has already been involved in food-seeking behaviors mostly associated to positive (motivation, reward, reinforcement) or negative valance (Labouèbe et al., 2016; Otis et al., 2017, 2019; Do-Monte et al., 2017; Wang et al., 2021a; Engelke et al., 2021) as well as in stress and emotional arousal (Hsu et al., 2014). However, whether and how the PVT and its afferent CA inputs may contribute to the regulation of food intake and metabolic efficiency in stress-related contexts remain to be fully established.
In order to dissect the contribution of PVT CA inputs in the regulation of food intake, we suppressed local CA inputs by microinjecting the neurotoxin 6-OHDA into the PVT and performed several experiments aimed at revealing the adaptive strategies associated to the regulation of feeding and energy balance. Here, we demonstrate that CA inputs to the PVT exerted a modulatory role on food intake and metabolic efficiency specifically in stress-related contexts since no major alterations were detected in basal conditions. Notably, we found that 6-OHDAPVT-lesioned mice, following both exteroceptive (environmental) and interoceptive (body energy) stressors, showed enhanced food intake and metabolic efficiency.
Altogether, our results reveal a novel neuronal network by which stressors impinge on the regulatory allostatic processes underlying feeding behaviors, metabolic efficiency and nutrient partitioning to scale stress-associated adaptive responses.
Materials and methods
Ethics statement and animals
All experiments were approved by the Animal Care Committee of the Université Paris Cité (APAFiS #35585 and #11003) and carried out following the 2010/63/EU directive. 8-14 weeks old male C57Bl/6J mice (Janvier, Le Genest St Isle, France) were used and housed in a room maintained at 22 +/-1 °C, with a light period from 7h00 to 19h00. Regular chow diet (3.24 kcal/g, reference SAFE® A04, Augy, France) and water were provided ad libitum unless otherwise stated.
Stereotaxic microinjections for viral tracing studies and 6-OHDA-induced catecholaminergic denervation
Mice were anaesthetized with isoflurane (3.5% for induction, 1.5% for maintenance), administered with Buprécare® (buprenorphine 0.3 mg) and Ketofen® (ketoprofen 100 mg), and placed on a stereotactic frame (Model 940, David Kopf Instruments). During surgery, body temperature was maintained at 37°C.
6-OHDA-HCl (Sigma-Aldrich, #H4381) was dissolved in a saline solution (NaCl 0.9% w/v) containing 0.02% of ascorbic acid at a final concentration of 6 µg/µl. Animals were randomly assigned to either 6-OHDA or vehicle microinjections. A volume of 0.35 µl of 6-OHDA or vehicle (0.02% ascorbic acid) was injected into the PVT (L= 0.0; AP= -1.46; V= -2.8, mm) at an infusion rate of 0.05 µl/min. The injection needle was carefully removed after waiting 5 minutes at the injection site. 24-hrs after, animals were re-administered with Buprécare® and Ketofen®, and had facilitated access to jelly food (DietGel Boost #72-04-5022, Clear H2O) for 2 consecutive days. Recovery from surgery was monitored during 3-5 days post-surgery. Animals were allowed to recover for 3-4 weeks before any experimental evaluation.
pAAV-CAG-tdTomato (titer ≥ 1×10¹³ vg/mL) was a gift from Edward Boyden (Addgene viral prep #59462-AAV9; http://n2t.net/addgene:59462; RRID:Addgene_59462). A volume of 0.20 µl of pAAV-CAG-tdTomato was injected into the PVT (L= 0.0; AP= -1.46; V= -2.8, mm) at an infusion rate of 0.05 µl/min. The injection needle was carefully removed after waiting 5 minutes at the injection site. Viral expression was evaluated 4 weeks after microinjection.
Indirect calorimetry and metabolic efficiency analysis
Indirect calorimetry was performed as previously described (Berland et al., 2022). Mice were monitored for whole energy expenditure (EE), O2 consumption and CO2 production, respiratory exchange rate (RER=VCO2/VO2), fatty acid oxidation (FAO), and locomotor activity using calorimetric cages with bedding, food and water (Labmaster, TSE Systems GmbH, Bad Homburg, Germany). Ratio of gases was determined through an indirect open circuit calorimeter (Arch et al., 2006; Even & Nadkarni, 2012). This system monitors O2 and CO2 concentration by volume at the inlet ports of a tide cage through which a known flow of air is being ventilated (0.4 L/min) and compared regularly to a reference empty cage. For optimal analysis, the flow rate was adjusted according to the animal body weights to set the differential in the composition of the expired gases between 0.4-0.9% (Labmaster, TSE Systems GmbH, Bad Homburg, Germany). The flow was previously calibrated with O2 and CO2 mixture of known concentrations (Air Liquide, S.A. France). O2 consumption and CO2 production were recorded every 15 min for each animal during the entire experiment. Whole energy expenditure (EE) was calculated using the Weir equation for respiratory gas exchange measurements. Food consumption was measured as the instrument combines a set of highly sensitive feeding sensors for automated online measurements. Mice had access to food and water ad libitum. To allow measurement of every ambulatory movement, each cage was embedded in a frame with an infrared light beam-based activity monitoring system with online measurement at 100 Hz. The sensors for gases and detection of movements operated efficiently in both light and dark phases, allowing continuous recording. When required, the inversion of circadian light/dark cycles was programmed using the Labmaster software.
Body mass composition was analyzed using an Echo Medical systems’ EchoMRI (Whole Body Composition Analyzers, EchoMRI, Houston, USA), according to manufacturer’s instructions. Readings of body composition were given within 1 min. Data analysis was performed on Excel XP using extracted raw values of VO2 consumed (expressed in ml/h), VCO2 production (expressed in ml/h), and energy expenditure (kcal/h).
Novelty-suppressed feeding (NSF)
After an overnight fasting, mice were placed in a cage (40×40×40 cm) with a single regular chow pellet in the center. The latency (time in seconds) to eat was scored. To measure food intake, food consumption was evaluated 60 minutes after the beginning of the test.
Open field (OF)
Spontaneous exploratory behavior was monitored in an open field (40×40×40 cm, BIOSEB) for 20 min, video-tracked and analyzed using the SMART software (BIOSEB). The open field was wiped with 70% ethanol between sessions.
Infrared temperature measurements
Thermogenesis was visualized using a high-resolution infrared camera (FLIR E8; FLIR Systems, Portland, OR, USA). To measure the temperature (°C) of the brown adipose tissue (BAT, interscapular regions), lower back and tail, images were captured before and after the open field (OF) test. Infrared thermography images were analyzed using the FLIR TOOLS software.
Food intake following food or water deprivation
In two distinct experiments we measured food intake following either an overnight fasting (food deprivation) or water deprivation.
Overnight fasting
Mice were first weighted in the morning following an overnight fasting to ensure the loss of body weight. Then, they were exposed to pre-weighted chow pellets. Food intake was measured at the following time points: 30 min, 1h, 2h, 3h, 4h.
Overnight water deprivation
Mice were first weighted in the morning following an overnight water deprivation to ensure the loss of body weight. Then, they were exposed to a calibrated drinking bottle and pre-weighted chow pellets. Water consumption and food intake were measured at the following time points: 30 min, 1h, 2h, 3h, 4h.
Food intake induced by 2-DG and ghrelin
Mice were handled and injected with vehicle during 3 consecutive days before drugs administration. After this habituation period, they were administered with ghrelin (#1465, Tocris, 0.5 mg/kg, i.p.) or 2-DG (#14325, Cayman Chemical, 500 mg/kg, i.p.) and exposed to chow pellets 30 min after the injections. Food intake was measured during 3 hours. For 2-DG treated mice, blood glucose (counterregulatory response) was measured from the vein tail using a glucometer (Menarini Diagnotics, Rungis, France) at 0 and 30 min.
Glucose dynamics
Oral glucose tolerance test (OGTT)
Animals were fasted 5 hours before receiving a bolus of glucose solution (2 g/kg) by gavage. Blood glucose (hyperglycemia) was measured from the vein tail using a glucometer (Menarini Diagnotics, Rungis, France) at 0, 5, 10, 15, 30, 60, 90, and 120 min.
Insulin tolerance test (ITT)
Animals were fasted 5 hours before receiving an injection of insulin (0.5 U/kg, Novo Nordisk, i.p.). Blood glucose (hypoglycemia) was measured from the vein tail at 0, 5, 10, 15, 20, 30, 60, 90, and 120 min.
Tissue preparation and immunofluorescence
Animals were injected with pentobarbital (500 mg/kg, i.p., Sanofi-Aventis, France). Once anaesthetized, they were transcardially perfused with cold (4°C) PFA 4% for 5 minutes. Brains were collected, put overnight in PFA 4% and then stored in PBS (4°C). 40 µm-thick sections were sliced with a vibratome (Leica VT1000S, France) and stored in a cryoprotective solution at -20 °C until immunofluorescence investigations. Immunofluorescence on brain slices was performed as previously described (Gangarossa et al., 2019; Berland et al., 2020).
Briefly, sections were processed as it follows. Day 1: free-floating sections were rinsed in Tris-buffered saline (TBS; 0.25 M Tris and 0.5 M NaCl, pH 7.5), incubated for 5 min in TBS containing 3% H2O2 and 10% methanol, and then rinsed three times for 10 min each in TBS. After 15 min incubation in 0.2% Triton X-100 in TBS, sections were rinsed three times in TBS again. Slices were then incubated 48 hrs at 4°C with the following primary antibodies: rabbit anti-cFos (1:1000, Synaptic Systems, #226 003), rabbit anti-TH (1:1000, Millipore, #AB153) or rat anti-DAT (1:500, Millipore, #MAB369). Sections were rinsed three times for 10 min in TBS and incubated for 60 min with secondary donkey anti-rabbit Cy3 AffiniPure (1:1000, Jackson ImmunoResearch, #711-165-152) or donkey anti-rat Cy3 AffiniPure (1:1000, Jackson ImmunoResearch, #712-165-153). Sections were rinsed for 10 min twice in TBS, stained with DAPI (10 min) and rinsed in TB (0.25 M Tris) before mounting.
Acquisitions were performed with a confocal microscope (Zeiss LSM 510). The objective (10X) and the pinhole setting remained unchanged during the acquisition of a series for all images. Depending on the extension of the region of interest, either single or mosaic acquisitions were conducted. Quantification of immunoreactive cells (cFos- or TH-positive neurons) was performed using the cell counter plugin of the ImageJ software taking as standard reference a fixed threshold of fluorescence. For cell counting, three (TH) or two (cFos) rostro-caudal levels for each brain region were used. Quantifications of immunopositive neurons were averaged between hemispheres and then summed for consecutive brain slices.
Statistics
All data are presented as mean ± SD. Statistical tests were performed with Prism 7 (GraphPad Software, La Jolla, CA, USA). Detailed statistical analyses are listed in the Statistical Summary Table. Normality was assessed by the D’Agostino-Pearson test. Depending on the experimental design, data were analyzed using either Student t-test (paired or unpaired) with equal variances, One-way ANOVA or Two-way ANOVA. The significance threshold was automatically set at p<0.05. ANOVA analyses were followed by Bonferroni post hoc test for specific comparisons only when overall ANOVA revealed a significant difference (at least p<0.05).
Results
Catecholaminergic inputs modulate the activity of PVT-neurons
To study whether catecholaminergic (TH-positive) inputs to the PVT participate/contribute to the regulation of food intake, we decided to ablate catecholamine (CA) fibers projecting to the PVT by locally microinjecting 6-OHDA (Figure 1A), a neurotoxin reuptaken by DAT- and/or NET-expressing terminals. Since the PVT extends throughout the midline thalamic axis, we decided to mainly focus on the mid-posterior PVT as this region has been shown to be involved in stress and stress-induced hypophagia (Heydendael et al., 2011; Barson et al., 2020; Barrett et al., 2021). Indeed, 6-OHDA was able to strongly reduce TH immunostaining in the PVT (Figure 1B), with few remaining terminals most likely corresponding to NET-negative TH fibers (terminals releasing adrenaline and/or devoid of monoamine transporters) since the PVT does not seem to contain DAT-positive terminals as compared to DAT-rich regions such as the tail of the striatum [TS, (Gangarossa et al., 2013; Valjent & Gangarossa, 2021)] and the central amygdala (Suppl. Figure 1A, https://doi.org/10.6084/m9.figshare.19683228.v1).
We therefore examined the regional sources of our 6-OHDA-induced degeneration of TH-afferents. We focused on putative noradrenergic TH-positive projecting neurons since the PVT do not receive dopaminergic inputs from the midbrain (SNc and VTA) (Li et al., 2014) and it harbors only minor, if any, scattered DAT-fibers (García-Cabezas et al., 2009; Clark et al., 2017). Immunofluorescence analysis revealed a significant reduction of TH-neurons in the nucleus tractus solitarius (NTS) and the locus coeruleus (LC), in line with the presence of Slc6a2 (NET)-cathecolaminergic neurons in these nuclei (Schroeter et al., 2000; Zhang et al., 2021) and the sensitivity of these neurons to 6-OHDA (Szot et al., 2012b; Lin et al., 2013). However, we observed a more robust reduction of PVT-projecting TH-neurons in the NTS (−34.2%) compared to the LC (−15.1%) (Figure 1C, D). No differences were observed in the A1 area of the hindbrain (Figure 1C, D) as well as in the hypothalamus (Suppl. Figure 1B, https://doi.org/10.6084/m9.figshare.19683228.v1).
Then we investigated whether the reduction in local TH-afferents was followed by the modulation of PVT-neurons activity. Since the PVT shows higher activity during wakefulness (Ren et al., 2018), mice were perfused 1h before the onset of the dark phase (spontaneous feeding period) and at basal conditions. Using cFos as a molecular proxy of cellular activity, we observed a significant increase in cFos-positive cells in the PVT of 6-OHDAPVT compared to ShamPVT mice (Figure 1E, F). This set of results suggests that a subset of hindbrain CA inputs (hindbrainTH→PVT projections) may serve as modulators of PVT-neurons activity.
Catecholaminergic inputs to the PVT contribute to novelty-induced hypophagia
Following 3-4 weeks from 6-OHDA microinjections, no major differences in body weight were detected in 6-OHDAPVT compared to ShamPVT mice (group-housed animals, Figure 2A). In order to study the role of PVT catecholaminergic inputs in the regulation of feeding patterns, 6-OHDAPVT and ShamPVT mice were single-housed (Figure 2B). Environmental and social isolation, as triggered by single housing, represent behavioral/environmental stress-like allostatic stimuli (Lee et al., 2020, 2021) which can trigger a transient reduction of food intake (Takatsu-Coleman et al., 2013; Benfato et al., 2022). Interestingly, we observed that 6-OHDAPVT mice were less sensitive to single housing-induced hypophagia compared to ShamPVT mice (Figure 2C). This phenotype prompted us to investigate whether catecholaminergic inputs to the PVT were important in driving feeding patterns and metabolic adaptations to exposure to other behavioral/environmental challenges.
After 2 weeks of habituation to single-housing, we analyzed the metabolic efficiency of 6-OHDAPVT and ShamPVT mice exposed to a novel environment during two consecutive exposures during the dark period (spontaneous feeding period, Figure 2D). While ShamPVT mice transiently (Exp.1) showed a reduction in food intake during the dark period (Figure 2E, E1), 6-OHDAPVT mice were again less sensitive to novelty-induced hypophagia. No differences in food intake were measured during Exp.2 period between the two groups, indicating a rapid restoration of homeostatic regulations associated to environmental habitation (Figure 2E, E1). This phenotype was not associated to changes in locomotor activity as indicated by the similar patterns of exploration (Exp.1, novelty) and habituation (Exp.2) (Figure 2F, F1). We also measured key whole-body metabolic parameters such as the respiratory exchange ratio (RER, indicative of the energy substrates used, RER=∼1 for carbohydrates, RER=∼0.7 for lipids), fatty acid oxidation (FAO) and energy expenditure (EE) during the exposure to the novel environment (Exp.1). Compared to ShamPVT mice, 6-OHDAPVT animals showed increased RER (Figure 2G, G1) and decreased FAO (Figure 2H, H1), mirroring the changes in food intake and indicating a shift of energy substrates (carbohydrates and lipids) use during exposure to a novel environment. However, these adaptations did not impact on energy expenditure (Figure 2I, I1), suggesting that nutrient partitioning (Joly-Amado et al., 2012), rather than total energy balance, was affected by the loss of hindbrainTH→PVT fibers.
One may wonder whether the absence of novelty-induced hypophagia in 6-OHDAPVT mice may be associated to enhanced perception and/or reward value of food. To investigate this aspect, ShamPVT and 6-OHDAPVT mice were intermittently (1h/day) exposed to high-fat high-sugar (HFHS) diet during two consecutive days. No differences in HFHS food intake were observed between groups (Suppl. Figure 2A, https://doi.org/10.6084/m9.figshare.19683228.v1), suggesting intact food palatability, perception and preference. These results suggest that hindbrainTH→PVT fibers represent an important node for the integration of homeostatic regulations.
Exposure to novel environments can lead to the occurrence of anxiogenic traits and novelty-triggered thermogenic adaptations (Lecorps et al., 2016) which may impact on, and therefore confound, the mechanisms underlying feeding strategies. Thus, we performed an open field test (OF, Figure 3A) to evaluate both anxiety and thermogenic adaptations. We observed no significant differences in anxiety-like parameters between ShamPVT and 6-OHDAPVT animals (Figure 3B-F). Moreover, both groups showed similar thermogenic enhancements in the brown adipose tissue (BAT), the lower back and the tail (Figure 3G-J). These results indicate that PVT-projecting TH-afferents modulate feeding patterns and metabolic efficiency independently from affective (anxiety) and thermogenic adaptations.
Feeding patterns and metabolic efficiency highly depend on circadian rhythms and strong functional interactions exist between circadian rhythms, feeding and energy balance (Challet, 2019). Moreover, the PVT, which is bidirectionally connected with the suprachiasmatic nucleus (SCN) (Peng & Bentivoglio, 2004; Colavito et al., 2015), has been pointed as a contributor to circadian cycles and its activity varies depending on active/inactive phases (Colavito et al., 2015; Kirouac, 2015). Thus, we decided to investigate whether PVT catecholaminergic inputs participated to the adaptive metabolic strategies occurring during circadian challenges by inverting the light/dark cycle (Figure 4A). First, before inverting the light/dark cycle, we confirmed (Figure 2) that exposure to the new environment was associated to reduced novelty-induced hypophagia in 6-OHDAPVT compared to ShamPVT mice (Suppl. Figure 3A, B, https://doi.org/10.6084/m9.figshare.19683228.v1). Then, after habituation, light/dark cycles were inverted. When measuring food intake, we observed that both experimental groups rapidly shifted and adapted toward the new light/dark schedule (average of first 2 days of standard and inverted cycles) (Figure 4B, B1). In fact, despite the adaptive shift (Figure 4B, B1), no differences were observed in the cumulative food intake (Figure 4C). Moreover, we also detected a similar pattern of circadian adaptation in the EE profile of both groups (Figure 4D, D1), with an increased EE in the 7h-19h inverted period (iDark) and a decreased EE in the 19h-7h inverted period (iLight) (Figure 4E, E1). Interestingly, we found a significant difference in FAO and RER. In particular, while ShamPVT animals rapidly adapted to the inverted light/dark cycle (Figure 4F), 6-OHDAPVT mice did not show significant changes in FAO during the 7h-19h inverted period (iDark) (Figure 4F1, G), whereas both groups showed increased FAO during the 19h-7h inverted period (iLight) (Figure 4G1). In line with this shift in energy substrates partitioning, measurements of RER indicated an impaired adjustment of metabolic efficiency during the iDark period (7h-19h) in 6-OHDAPVT mice (Figure 4H, H1, I, I1).
These results suggest that hindbrainTH→PVT inputs, although not involved in the adaptation of food intake to circadian challenges, are important in adjusting peripheral energy substrates utilization (i.e lipids and carbohydrates as revealed by RER and FAO) and overall metabolic efficiency.
Catecholaminergic inputs to the PVT contribute to feeding under physiological and metabolic challenges
Next, we wondered whether hindbrainTH→PVT inputs guided feeding following physiological and metabolic stressors. First, to mimic a conflict between hunger and environmental stress, we decided to study food intake in overnight fasted mice in a novelty-induced hypophagia test. After an overnight fasting, both grouped showed a similar reduction in body weight and plasma glucose levels (Suppl. Figure 4A, B, https://doi.org/10.6084/m9.figshare.19683228.v1). Then mice underwent the novelty-suppressed feeding (NSF) test (Figure 5A). ShamPVT and 6-OHDAPVT mice showed similar latency to eat (Figure 5B), with a significant difference in food intake which was higher in 6-OHDAPVT mice (Figure 5C, D). The NSF test (latency to eat) is mainly used to assess anxiety- and depressive-like phenotypes. Therefore, our results suggest that the increased food intake observed following PVT catecholaminergic ablation may not be confounded by potential alterations triggered by anxiety. This is in line with our above-mentioned observations using the open field test (Figure 3A-F).
Then, we used an acute restraint (immobilization) paradigm which is known to alter metabolism and food intake (Rybkin et al., 1997; Vallès et al., 2000; Rabasa & Dickson, 2016). ShamPVT and 6-OHDAPVT mice underwent a 30 min acute restraint and food intake was measured during the dark period. Although both groups showed a significant reduction in food intake (Figure 5E), stress-induced hypophagia was significantly more pronounced in ShamPVT compared to 6-OHDAPVT mice (Figure 5E).
To further investigate the role of hindbrainTH→PVT inputs in scaling feeding, we used metabolic stressors to modulate food intake. First, ShamPVT and 6-OHDAPVT mice were administered with 2-deoxy-d-glucose (2-DG) which, in virtue of its glucoprivic effects (neuroglucopenia), elicits food consumption as well as the typical glucose counterregulatory response (CRR) (Pénicaud et al., 1986; Lewis et al., 2006). Moreover, PVT-neurons are highly sensitive to glucoprivation (Labouèbe et al., 2016). In this conditions, 6-OHDAPVT mice consumed more chow food than ShamPVT mice (Figure 6A), even though the magnitude of the glucose excursion as counterregulatory response was similar between groups (Figure 6B). This result suggests that hindbrainTH→PVT projections are required to fully express feeding adaptions to glucoprivic conditions but are dispensable for the autonomic control of glycogen breakdown and glucose production in CRR. In the same line, glucose clearance dynamics during an oral glucose tolerance test (OGTT, Figure 6C) or an insulin tolerance test (ITT, Figure 6D) were similar between sham and 6-OHDAPVT mice, indicating that glucose metabolism and insulin sensitivity remained unaltered following the loss of hindbrainTH→PVT inputs.
Second, we performed a fasting-refeeding test to mimic a negative energy balance (food deprivation). In line with the NSF test (Figure 5C, D), after an overnight fasting and a similar loss of body weight (Suppl. Figure 4C, https://doi.org/10.6084/m9.figshare.19683228.v1), both experimental groups showed an enhanced food intake with 6-OHDAPVT mice consuming more food than ShamPVT mice (Figure 6E). Third, we also decided to measure drinking and food intake in overnight water-deprived animals. Both groups showed again a similar decrease in body weight (Suppl. Figure 4D, https://doi.org/10.6084/m9.figshare.19683228.v1). After deprivation, mice were exposed to water. While no differences were observed in drinking behavior (Figure 6F), 6-OHDAPVT mice again consumed more food than ShamPVT mice (Figure 6G). These results suggest that hindbrainTH→PVT inputs scale food intake also following physiological and metabolic stressors.
In order to assess whether orexigenic signals without metabolic challenges also required an intact PVT catecholaminergic transmission, we administered ghrelin in fed mice. As shown in Figure 6H, ghrelin similarly induced an increase in food intake in both groups, thereby indicating that canonical orexinergic circuits are not altered.
Reduced catecholaminergic transmission in the PVT promotes the activation of hypothalamic regions
The above-mentioned results point to hindbrainTH→PVT afferents as major actors in scaling food intake and metabolic efficiency. Since these homeostatic functions highly depend on the hypothalamus, classically described as the master regulator of energy balance (Dietrich & Horvath, 2013; Timper & Brüning, 2017), we decided to study whether 6-OHDAPVT mice were characterized by an altered basal activity (cFos-positive cells) of key hypothalamic regions such as the dorsomedial hypothalamus (DMH), the ventromedial hypothalamus (VMH), the lateral hypothalamus (LH) and the arcuate nucleus (Arc). As for Figure 1, mice were perfused 1h before the onset of the dark phase. Interestingly, in 6-OHDAPVT mice we observed an increase in cFos-positive cells specifically in the DMH and LH (Figure 7A-C) compared to ShamPVT mice, whereas no differences were detected in the VMH and Arc (Figure 7A, D, E).
In order to see whether PVT-neurons may potentially modulate hypothalamic functions by directly projecting to the DMH and LH, we microinjected an AAV9-CAG-TdTomato virus in the PVT (Figure 7F, G). As shown in Figure 7G, we observed direct PVT→DMH and PVT→LH projections. These results suggest that the increased activity of PVT-neurons following catecholamines depletion (Figure 1) may impact, directly (Figure 7G) and/or indirectly (polysynaptic circuits), on the regulatory activity of the hypothalamus which may ultimately result in the modulation of food intake and metabolic efficiency.
Discussion
The regulation of food intake represents one of the most complex biological functions. Pivotal for adaptation and survival, the regulatory processes underlying feeding are constantly shaped by signals reflecting/sensing physiological adjustments. In virtue of their heterogeneity (exteroceptive vs interoceptive sources), stress-like allostatic stimuli are indeed powerful modulators of food intake, feeding habits and metabolic adaptations.
In this study, we report that hindbrain catecholaminergic (putative noradrenergic) inputs to the PVT play a key role in modulating food intake and metabolic efficiency in stress-related contexts. In fact, permanent ablation of TH-afferents to the PVT resulted in enhanced food intake, adjusted metabolic efficiency and nutrient partitioning whenever environmental, behavioral, physiological and/or metabolic (acute/transient) stressors were introduced as dependent variables of feeding behaviors. In particular, 6-OHDAPVT mice were resistant or less sensitive to environmental and behavioral stress-induced hypophagia and showed enhanced feeding patterns following physiological and metabolic challenges. The different nature of stressors used in this study highlights the highly conserved role of CA inputs to the PVT in readily scaling feeding and metabolic adaptations. Moreover, beyond the impact on feeding and metabolic efficiency, it is important to mention that stressors-elicited homeostatic adaptations such as energy expenditure and thermogenesis did not depend on PVT CA inputs, suggesting a functional tropism of hindbrainTH→PVT circuits toward food intake on one hand and peripheral nutrient partitioning on the other. This is relevant since previous studies have suggested that PVT-neurons, by facilitating hypothalamic-pituitary-adrenal (HPA) responses (Bhatnagar et al., 2000), may contribute to the regulation of core temperature rhythms as well as body weight gain in chronically stressed rats (Bhatnagar & Dallman, 1999). Whether distinct PVT networks (inputs/outputs) are differently engaged by acute vs chronic stressors on the regulation of body hemostasis will deserve in-depth investigations. Overall, these results indicate that the PVT, a key node of the limbic circuitry (Barson et al., 2020), contributes to the elaboration of food-related decisions and strategies by integrating, among others, also catecholaminergic information. In addition, our results provide new evidence for the existence of distinct, but converging, hindbrain CA inputs (a subset of NTSTH- and LCTH-neurons) capable of gating food-related homeostatic adaptations under transient stress-like allostatic stimuli.
Surprisingly, we found that the NTS represented one of the major sources of TH-positive projections to the PVT. In fact, local microinjection of 6-OHDA resulted in a significant reduction of TH-expressing neurons in the NTS and to a lesser extend in the LC, sparing CA neurons in the medulla (A1 area) and the hypothalamus, the latter known to send only minor scattered projections to the PVT (Wang et al., 2021b). Although to our knowledge no other studies have functionally assessed this NTSTH→PVT connection, the impact of 6-OHDAPVT on NTSTH-neurons is in line with the presence of a dense plexus of PVT-reaching TH fibers when fluorescent recombinant markers are directly microinjected in the NTS of Th-Cre animals (Aklan et al., 2020) as well as with a recent retrograde viral study identifying a subset of NTSTH-neurons projecting to the PVT (Kirouac et al., 2022). This evidence is of paramount important since NTS and LC catecholaminergic neurons, by converging onto the PVT, may synergistically modulate feeding patterns and metabolic efficiency in stressogenic contexts. In fact, NTSTH- and LCTH-neurons are well-known to modulate food intake and stress/novelty, respectively (McCall et al., 2015; Roman et al., 2016; Takeuchi et al., 2016).
Recent studies have shown that activation of CA-releasing NTSDBH/TH-neurons may result in a reduction (Roman et al., 2016; Chen et al., 2020) as well as in an increase (Aklan et al., 2020; Chen et al., 2020) of food intake depending on CA cell types and/or projection sites [parabrachial nucleus (PBN) vs arcuate nucleus (Arc)]. Although not directly assessed in our study, our results suggest that NTSTH→PVT projecting neurons may serve as anorexigenic stimuli since their ablation enhances food intake under stress-related contexts. Indeed, it may be legitimately argued that the use of local microinjections of 6-OHDA may result in the degeneration of LCTH- and NTSTH-neurons projecting to the PVT but eventually to also other brain regions. However, loss of hindbrainTH neurons projecting to the medial hypothalamus resulted in a loss of glucoprivation-induced feeding (Fraley & Ritter, 2003; Hudson & Ritter, 2004), while in our study we show an increased food consumption under glucoprivic conditions when hindbrainTH→PVT projections were ablated. Moreover, opto-activation of NTSTH→Arc projections leads to an increase in food intake (Aklan et al., 2020), whereas in our case enhanced food intake was elicited by the absence of hindbrainTH→PVT projections. These effects may substantiate the functional selectivity of hindbrainTH→PVT projections in the adaptive responses of feeding, metabolic efficiency and nutrient partitioning to stress-related contexts. Indeed, future investigations using projection-specific optogenetics and/or chemogenetics may definitely help in dissecting out the distinct and, most importantly, synergistic roles of NTSTH→PVT and LCTH→PVT transmissions in guiding the tight interaction between food intake, metabolism and stress-like allostatic stimuli.
Moreover, we observed that deletion of hindbrainTH→PVT inputs led to an increase in PVT-cells activity (cFos), suggesting that catecholamines may act, directly and/or indirectly, as negative modulators onto PVT-neurons. At first, this is surprising and counterintuitive since ex vivo bath-applications of DA (precursor of NE) or NE, which can be synaptically co-released from CA terminals (Smith & Greene, 2012; Kempadoo et al., 2016), lead to disinhibition [DA, (Beas et al., 2018)] or activation [NE, (Wang et al., 2021b)] of PVT-neurons. However, it should be mentioned that mid-posterior PVT-neurons express DA D2 and D3 receptors (Rieck et al., 2004; Clark et al., 2017; Beas et al., 2018; Gao et al., 2020) as well as NE receptors such as the α1, α2, β1 and β2 receptors (Rainbow et al., 1984; Pieribone et al., 1994; Rosin et al., 1996). Indeed, (i) how this variety of G-protein-coupled receptors (Gi-, Gs-, Gq- and β-arrestin-coupled receptors) mechanistically contribute to the overall modulation of PVT-neurons and (ii) whether 6-OHDA-induced TH deletion reorganizes the expression of the above-mentioned CA receptors in the PVT require future investigations. Although our results cannot distinguish between the functional roles of distinct catecholamines onto their associated multiple receptors located onto PVT-neurons, it is worth to mention that the PVT does not receive pure DA-fibers from the midbrain (SNc and VTA) (Li et al., 2014; Papathanou et al., 2019) and that direct 6-OHDA-induced LCTH-neurons loss was associated to an increased expression of Gq-coupled α1 receptor in the thalamus (Szot et al., 2012a) which may explain, at least in part, the increase in PVT cFos-neurons.
The enhanced activation of PVT-neurons in 6-OHDAPVT mice and the associated feeding behaviors are in line with reports showing that stressors as well as hunger are able to activate PVT-neurons (Bubser & Deutch, 1999; Beas et al., 2018; Hua et al., 2018). In addition, our results are also in line with a recent report showing that activation of PVT-neurons by oxytocin was efficient in suppressing stress-induced hypophagia (Barrett et al., 2021). However, it is worth to mention that satietogenic signals are also able to activate PVT-neurons (Ong et al., 2017), therefore indicating that PVT excitatory (glutamate) neurons may be actually segregated into several cell types with distinct neurochemical, cellular and functional features. This is already supported by the existence of at least two neuronal populations [galanin- and dopamine 2 receptor (D2R)-positive neurons, (Gao et al., 2020)] and, as already suggested by the presence of several neuropeptides in PVT-neurons (Curtis et al., 2020), it may not be hazardous to hypothesize that future cell type-specific transcriptomic analyses will reveal new sub-families and clusters.
The homeostatic processes underlying food intake, energy balance and metabolic efficiency strongly depend on the activity of the hypothalamus (Dietrich & Horvath, 2013; Timper & Brüning, 2017). We observed that depletion of TH-afferents to the PVT resulted not only in the activation of PVT-neurons but also in the concomitant activation of hypothalamic regions, notably the lateral (LH) and the dorsomedial (DMH) hypothalamus. Indeed, activation of LH and DMH cell types has been shown to promote feeding (Jennings et al., 2015; Navarro et al., 2016; Otgon-Uul et al., 2016; Jeong et al., 2017) even following anxiogenic environmental cues (Cassidy et al., 2019). Although we cannot rule out yet whether and how the adaptive activation of PVT-neurons following TH deletion may be responsible for the direct activation of LH and DMH regions, it is interesting to note that PVT excitatory (glutamate) neurons also project to the hypothalamus (Engelke et al., 2021; Li et al., 2021), therefore potentially modulating feeding and energy homeostasis. This is also supported by our viral tracing strategy which revealed direct PVT→DMH/LH projections. However, we cannot formally exclude that the partial loss of hindbrain TH-neurons may impact on the hypothalamic activity in virtue of other circuits (hindbrainTH→hypothalamus and/or hindbrainTH→PBN→hypothalamus paths). Indeed, while the existence of a hypothalamus→PVT→accumbal circuit seems critical for behavioral adaptations (Betley et al., 2013; Zhang & van den Pol, 2017; Otis et al., 2019; Meffre et al., 2019; Zhang et al., 2020; Iglesias & Flagel, 2021; Engelke et al., 2021), our results, together with previous and recent literature (Otake et al., 1994; Ong et al., 2017; Beas et al., 2018; Sofia Beas et al., 2020; Li et al., 2021), also suggest a hindbrain→PVT→hypothalamus path that may regulate homeostatic functions requiring the integration of exteroceptive and interoceptive signals.
In conclusion, the PVT has been classically positioned as a functional node of the limbic circuit (Barson et al., 2020). Only recently the hypothesis of the PVT as a homeostatic relay has been proposed (Penzo & Gao, 2021). Altogether, our results support the working hypothesis according to which the PVT, through its afferent connections with NTSTH- and LCTH-neurons, may represent a functional interface between homeostatic and emotional states, thereby leading to allostatic adaptations. This study, besides highlighting the existence of a dual hindbrain-to-thalamus connection, (i) provides new evidence to better understand the dynamic processes underlying the regulation of food intake and energy metabolism, and (ii) may serve as starting step to explore the functional relationships and comorbidities between psychiatric (stress) and metabolic (anorexia, obesity, binge eating) disorders.
Author Contributions
C.D. performed and analyzed most of the experiments. G.L. performed immunofluorescence studies. J.C. performed surgeries. S.L. provided critical feedback. G.G. conceived and supervised the whole project, and wrote the manuscript with contribution from all coauthors.
Data availability statement
All data are presented in the manuscript or supplementary information. For Suppl. Figures see https://doi.org/10.6084/m9.figshare.19683228.v1.
Competing interests
The authors declare no competing interests.
Figure legends
Suppl. Figure 1: Detection of DAT-positive fibers in the PVT and TH-positive neurons in the hypothalamus following PVT 6-OHDA microinjections. (A) Immunofluorescence detection of the dopamine transporter (DAT, red) and DAPI (blue) in the PVT, tail of the striatum (TS) and central amygdala (CeA). The lack of DAT-positive fibers in the PVT suggests no direct projections from dopamine-(DAT)-containing midbrain regions. Scale bar: 150 µm. (B) Immunofluorescence detection of TH (red) and DAPI (blue) in the hypothalamus of ShamPVT and 6-OHDAPVT mice. Scale bar: 500 µm.
Suppl. Figure 2: HindbrainTH→PVT inputs dos not alter palatability for HFHS diet. (A) Food intake in ShamPVT and 6-OHDAPVT mice following time-locked feeding (1h) of high-fat high-sugar (HFHS) diet during two consecutive days. Statistics: ** p<0.01, Day2 vs Day1 (ShamPVT mice); ## p<0.01, Day2 vs Day1 (6-OHDAPVT mice). No differences between experimental groups. Data are presented as mean ± SD. For statistical details see Statistical Summary Table.
Suppl. Figure 3: Confirmation of sensitivity to novelty-induced hypophagia in ShamPVT and 6-OHDAPVT mice before the inverted cycle. (A) Cumulative food intake during the dark phase (spontaneous eating) in ShamPVT and 6-OHDAPVT mice during the first exposure to a novel environment (calorimetric chambers). (B) Total food intake during the dark phase. Statistics: *** p<0.001, ** p<0.01, 6-OHDAPVT vs ShamPVT mice. Data are presented as mean ± SD. For statistical details see Statistical Summary Table.
Suppl. Figure 4: Effect of fasting and water deprivation in ShamPVT and 6-OHDAPVT mice. (A) Loss of body weight and (B) glucose variations in overnight fasted animals used for the novelty-suppressed feeding (NSF) test. (C) Loss of body weight in overnight fasted animals before the refeeding schedule. (D) Loss of body weight in overnight water-deprived animals before having access to water and chow pellets. No differences between experimental groups. Data are presented as mean ± SD. For statistical details see Statistical Summary Table.
Acknowledgments
We thank Olja Kacanski for administrative support; Isabelle Le Parco, Ludovic Maingault, Angélique Dauvin, Aurélie Djemat, Florianne Michel and Daniel Quintas for animals’ care; Benoit Bertrand for technical help. We acknowledge the Functional and Physiological Exploration platform (FPE) of the Université de Paris (BFA, UMR 8251) and the animal facility Buffon of the Institut Jacques Monod. This work was supported by the Fyssen Foundation, Nutricia Research Foundation, Allen Foundation Inc., Agence Nationale de la Recherche (ANR-21-CE14-0021-01), Fédération pour la Recherche sur le Cerveau and Association France Parkinson, Université Paris Cité and CNRS. G.L. was supported by the China Scholarship Council (CSC) fellowship.