We have previously shown that NTS^Lepr and NTS^Calcr cells represent distinct populations of NTS neurons (Ludwig et al., 2021a; Cheng et al., 2020a), and that the activation of either population mediates the non-aversive suppression of food intake (Cheng et al., 2020a; Cheng et al., 2020b). To determine the relative ability of each of these cell types to mediate the suppression of food intake and to determine whether they might play cumulative roles in the suppression of feeding, we injected an adeno-associated virus (AAV) to cre-dependently express the activating (hM3Dq) designer receptor exclusively activated by designer drugs (DREADD; AAV^FLEX-hM3Dq Sternson and Roth, 2014; Zhu and Roth, 2014) into the NTS of Lepr^Cre, Calcr^Cre, or dual Lepr^Cre;Calcr^Cre mice (producing Lepr^Dq, Calcr^Dq, and LC^Dq mice, respectively). We then examined the effects of CNO-mediated hM3Dq activation on food intake and related parameters in these mice (Figure 1). As expected, CNO treatment promoted FOS-immunoreactivity (-IR) in the NTS of Lepr^Dq, Calcr^Dq and LC^Dq mice, with the LC^Dq mice displaying increased FOS-IR compared to animals in which we had transduced only one population (Figure 1A).

As we showed previously (Cheng et al., 2020a; Cheng et al., 2020b), CNO almost completely abrogated food intake at the onset of the dark cycle in chow-fed Lepr^Dq and Calcr^Dq mice; LC^Dq mice responded similarly (Figure 1B). The same was true for mice fed high-fat diet (HFD) at the onset at the dark cycle, although the activation of NTS^Lepr neurons tended to suppress feeding less than NTS^Calcr and NTS^LC neurons at the 4 hr time point in this assay (Figure 1C). Following an overnight fast, the activation of NTS^Lepr neurons significantly decreased refeeding over 6 hr; the activation of NTS^Calcr cells alone decreased refeeding further than NTS^Lepr neurons, and the activation of NTS^LC neurons decreased refeeding even further- by approximately 90% (Figure 1D).

Similarly, activating NTS^Lepr neurons decreased feeding and body weight over a two-day treatment, activating NTS^Calcr neurons suppressed these parameters more effectively, and activating both cell types in LC^Dq mice reduced feeding by >90% for the first 24 hr and continued to reduce food intake compared to the individual cell types during the second 24 hr of the experiment (Figure 1E and F). While we had intended to continue treatment for another day, we discontinued CNO treatment after 48 hr because of animal welfare concerns due to the amount of weight loss exhibited by the LC^Dq mice. Hence, NTS^Calcr neurons suppress feeding more robustly than NTS^Lepr neurons, and NTS^Calcr and NTS^Lepr neurons suppress feeding cumulatively.

To determine whether the dramatic decrease in feeding promoted by NTS^LC neuron activation provoked avoidance responses, we examined whether CNO treatment in LC^Dq mice would provoke a CTA. We found that, although the gut malaise associated with LiCl injection promoted a robust CTA to either the novel exposure to HFD or saccharine in drinking water, NTS^LC activation failed to provoke such a response (Figure 1G and H). Thus, the coordinated activation of two non-aversive cell types (NTS^Lepr and NTS^Calcr Cheng et al., 2020a; Cheng et al., 2020b; Gaykema et al., 2017) fails to promote a CTA despite almost abolishing food intake.

We also expressed the inhibitory (hM4Di) DREADD in NTS^LC cells to examine the effect of acutely inhibiting these cells on short-term food intake (Figure 1—figure supplement 1). While CNO failed to significantly increase food intake during the first four hours of the dark cycle, stimulating hM4Di in NTS^LC cells increased food intake during refeeding with chow or HFD following an overnight fast. Thus, NTS^LC neurons contribute to the restraint of food intake during refeeding.

We previously showed that the long-term tetanus toxin (TetTox)-mediated silencing of NTS^Calcr neurons promotes increased food intake and body weight in animals exposed to HFD (Cheng et al., 2020a). To determine whether silencing NTS^Lepr cells might mediate similar effects (thereby implicating this additional NTS cell type in the control of long-term energy balance), we injected an AAV to cre-dependently express TetTox into the NTS of Lepr^Cre mice (AAV^FLEX-TetTox Kim et al., 2009; Lepr^TetTox mice) and examined their food intake and body weight during 6 weeks of chow feeding followed by an additional 6 weeks of HFD feeding (Figure 2A–D). While we detected no significant increase in food intake or body weight during chow feeding for Lepr^TetTox mice, these animals displayed a modest increase in food intake on HFD and gained more body weight over 6 weeks than control animals. Thus, NTS^Lepr neurons (like NTS^Calcr cells) contribute to the long-term control of energy balance.

To determine whether the additional silencing of NTS^Calcr neurons might exacerbate the increased food intake and body weight displayed by Lepr^TetTox mice, we used AAV^FLEX-TetTox to cre-dependently express TetTox in the NTS of Lepr^Cre;Calcr^Cre animals (LC^TetTox mice) (Figure 2E–H). While we detected no significant increase in food intake or body weight during 6 weeks of chow feeding in LC^TetTox mice compared to controls, LC^TetTox mice ate significantly more food and gained (~10 grams) more body weight than controls when exposed to HFD for 6 weeks.

Continuous analysis of feeding demonstrated increased food intake by LC^TetTox mice in the dark (active) cycle, consistent with decreased suppression of food intake in response to normal feeding (Figure 2—figure supplement 1). While comparing results among different strains of mice studied at different times requires cautious interpretation, these findings suggest that the silencing of multiple NTS neuron populations (as in LC^TetTox mice) increases feeding and weight gain relative to silencing single populations (NTS^Lepr in the current study, as well as previously-studied NTS^Calcr cells previously Cheng et al., 2020a; Figure 2I).

Given the dramatic effects on food intake and body weight observed upon silencing NTS^LC cells, we decided to test the requirement for the function of these cells in the weight loss response to VSG. We thus generated HFD-fed control and LC^TetTox animals as in Figure 2 and subjected them to VSG (Figure 2—figure supplement 1). While, as expected, HFD-fed LC^TetTox mice weighed more than HFD-fed control animals at the time of surgery, VSG also promoted more weight loss in these animals than in controls (both in terms of total weight lost and percent body weight lost) and similarly improved measures of glycemic control. Thus, NTS^LC neurons are not required for VSG-mediated weight loss, and other cell types must contribute to the suppression of food intake following VSG.

While NTS^Lepr and NTS^Calcr cells both mediate the non-aversive suppression of food intake, activation of NTS^Cck neurons promotes avoidance (e.g. CTA) responses, as well as decreasing feeding (Cheng et al., 2020a; Roman et al., 2017; Cheng et al., 2020b; Roman et al., 2016). Thus, to understand potential roles for avoidance-provoking NTS^Cck cells in the control of energy balance, we injected AAV^FLEX-TetTox to mediate the cre-dependent expression of tetanus toxin (TetTox) into the NTS of Cck^Cre mice (Cck^TetTox mice; Figure 3A) and examined their food intake and body weight during 6 weeks of chow feeding followed by an additional 6 weeks of HFD feeding (Figure 3B–D). While we detected no significant increase in food intake or body weight during chow feeding for Cck^TetTox mice, these animals gained more body weight over 6 weeks than control animals during HFD feeding. Thus, while it is possible that the inaccuracies inherent to food intake measurement prevented us from detecting an alteration in food intake in the Cck^TetTox animals, silencing NTS^Cck neurons might impact body weight by other means, including by altering calorie absorption or energy expenditure. In any case, NTS^Cck neurons, like NTS^Calcr and NTS^Lepr cells, contribute to the long-term control of energy balance.

To determine whether avoidance-provoking NTS^Cck neurons might play a role in the weight loss response to VSG, we decided to test whether the function of these cells for the weight loss response to VSG. We thus generated HFD-fed control and Cck^TetTox animals as in Figure 3 and subjected them to VSG (Figure 3—figure supplement 1). While, as expected, HFD-fed Cck^TetTox mice weighed more than HFD-fed control animals at the time of surgery, VSG promoted similar weight loss in these animals as in controls.

To understand potential cumulative effects on food intake for avoidance-provoking and non-aversive NTS cell types, we chose to examine the effects of manipulating NTS^Cck neurons (Cheng et al., 2020a; Roman et al., 2017; Roman et al., 2016) in combination with NTS^LC cells (collectively, NTS^LCK cells; Figures 4–5). We began by injecting AAV^FLEX-hM3Dq (Sternson and Roth, 2014; Zhu and Roth, 2014) into the NTS of Lepr^Cre;Calcr^Cre;Cck^Cre mice (producing LCK^Dq mice; Figure 4). We then examined the effects of CNO-mediated activation of NTS^LCK cells on food intake and related parameters in these animals.

As expected, CNO treatment promoted FOS-IR in the NTS of LCK^Dq mice (Figure 4A). As for LC^Dq mice (Figure 1), CNO almost completely abrogated food intake at the onset of the dark cycle in chow-fed and HFD-fed LCK^Dq mice, and strongly suppressed refeeding following an overnight fast (Figure 4B–D). We also assessed the effect of activating NTS^LCK neurons over a two-day treatment. CNO treatment dramatically decreased food intake in LCK^Dq mice and resulted in even larger decreases in body weight that in LC^Dq mice (Figures 1E–F–4E–F), confirming the effectiveness of the coordinated action of NTS^LCK neurons for the control of food intake and body weight.

Because food intake during CNO treatment was similarly low for LC^Dq and LCK^Dq mice (it is not possible to decrease food intake below zero), while the LCK^Dq mice lost more body weight, it is possible that the activation of NTS^Cck neurons promotes negative energy balance by altering energy absorption or expenditure as well as by decreasing feeding. Indeed, silencing NTS^Cck cells increased body weight without measurably altering food intake (Figure 3). We also found that, unlike NTS^LC cells, activating NTS^LCK cells provoked a strong CTA to HFD or saccharine in water, consistent with the avoidance-provoking nature of NTS^Cck neurons and the dominant nature of avoidance-provoking over non-avoidance-provoking hindbrain signals (Figure 4G–H).

To determine whether silencing NTS^Cck neurons might exacerbate the increased food intake and body weight displayed by LC^TetTox mice (and vice-versa), we injected AAV^FLEX-TetTox into the NTS of Lepr^Cre;Calcr^Cre;Cck^Cre animals (LCK^TetTox mice) (Figure 5A). While we had not detected increased food intake on chow diet in Lepr^TetTox, Calcr^TetTox, Cck^TetTox, or LC^TetTox animals (Figures 2–3), LCK^TetTox mice displayed a significant increase in food intake on chow, as well as on HFD (Figure 5B–C). Although LCK^TetTox mice also tended to gain more weight on chow diet, the effect was not statistically significant. However, these animals became significantly heavier than controls after the first week on HFD, and continued to gain weight relative to controls with each additional week (Figure 5D).

While one must exercise caution when comparing results among different strains of mice, we observed increased HFD food intake for LCK^TetTox animals compared to Cck^TetTox and LC^TetTox mice (Figure 5E). Similarly, LCK^TetTox mice tended to gain more weight on HFD compared to LC^TetTox and Cck^TetTox animals (Figure 5F). These findings suggest that simultaneously silencing multiple food intake controlling NTS neuron populations cumulatively increases food intake and body weight, irrespective of whether or not the neurons mediate avoidance responses.

To better understand the potential mechanisms underlying the cumulative nature of signals from the various NTS cell types, we treated control (wild-type), Lepr^Dq, Calcr^Dq, Cck^Dq, LC^Dq, and LCK^Dq mice with CNO and examined FOS-IR in the AP, NTS, parabrachial nucleus (PBN), and dorsomedial hypothalamic nucleus (DMH) (Figure 6—figure supplement 1). In addition to revealing the expected increase in NTS FOS-IR with the activation of each additional cell type (Figure 6A and D), we also observed differences in the magnitude and distribution of FOS-IR between cell types in other areas in a manner that was roughly additive. While CNO provoked little AP FOS-IR in Lepr^Dq mice, the activation of NTS^Calcr or NTS^Cck neurons tended to increase AP FOS-IR (Figure 6A and C). Furthermore, while CNO treatment in LC^Dq mice did not increase AP FOS-IR compared to that observed in Calcr^Dq mice, LCK^Dq demonstrated the greatest amount of AP FOS-IR.

In the PBN (Figure 6B and D), the LC^Dq mice demonstrated increased FOS-IR compared to Lepr^Dq or Calcr^Dq alone, although neither of these induced the colocalization of FOS-IR with CGRP (the activation of which is known to promote avoidance behavior Carter et al., 2015). In contrast, the activation of NTS^Cck cells alone or in combination with the other cell types provoked FOS/CGRP colocalization, consistent with the avoidance response provoked by any stimulus that includes the activation of NTS^Cck cells. CNO treatment in LCK^Dq mice provoked increased PBN FOS-IR compared to LC^Dq mice, suggesting the existence of substantial numbers of distinct PBN targets for NTS^Cck compared to NTS^LC neurons. The activation of DMH FOS-IR similarly increased with the stimulation of additional NTS cell types (Figure 6—figure supplement 1). Together, these findings suggest that each NTS population targets at least partially distinct downstream neurons in several brain regions. Thus, the activation of additional downstream targets by the stimulation of more populations of NTS neurons likely underlies the cumulative effects of activating multiple NTS cell types.

Given the dramatic effects on food intake and body weight observed upon silencing NTS^LCK cells, we decided to test the requirement for the combined function of these cells for the weight loss response to VSG (Figure 6—figure supplement 2). While, as expected, HFD-fed LCK^TetTox mice weighed more than HFD-fed control animals at the time of surgery, VSG promoted more body and fat mass loss in these animals than in controls. Food intake was decreased by VSG in LCK^TetTox animals, although it remained higher than for control mice that had received VSG, presumably reflecting the increased food intake of these mice at baseline. Thus, VSG normalized body weight and fat mass in LCK^TetTox mice and promoted similar improvements in indices of glucose homeostasis in LCK^TetTox and control mice. These findings suggest that, despite the important roles for these cells in the physiologic restraint of feeding and body weight, neurons other than NTS^LCK cells must mediate the response to VSG.

To determine whether additional (non-LCK) NTS neurons might contribute to the refeeding response in control animals and/or in mice subjected to VSG, we quantified the FOS response to refeeding in control or VSG animals expressing reporter fluorophores in NTS^LCK cells (Figure 6—figure supplement 3). This analysis revealed that NTS^LCK cells represent approximately 20% of feeding-activated NTS neurons in control mice, and approximately 35% of such neurons following VSG. Thus, in addition to neurons in the AP, hypothalamus, and elsewhere, feeding stimulates non-LCK neurons in the NTS and these feeding-activated non-LCK cells presumably contribute to food intake suppression and body weight control.

Mice were bred in our colony in the Unit for Laboratory Animal Medicine at the University of Michigan; these mice and the procedures performed were approved by the University of Michigan Committee on the Use and Care of Animals (Protocol# 00011066) and in accordance with Association for the Assessment and Approval of Laboratory Animal Care and National Institutes of Health guidelines. Mice were provided with food and water ad libitum (except as noted below) in temperature-controlled rooms on a 12 hr light-dark cycle. For all studies, animals were processed in the order of their ear tag number, which was randomly assigned at the time of tailing (before genotyping). ARRIVE guidelines were followed; animals were group-housed except for feeding and CTA studies.

We purchased male and female C57BL/6 mice for experiments and breeding from Jackson Laboratories. Lepr^cre and Calcr^cre mice have been described previously (Patterson et al., 2011; Pan et al., 2018) and were propagated by intercrossing homozygous mice of the same genotype. Cck^cre mice (Jax stock No.: 012706) for breeding were purchased from Jackson Labs (Bar Harbor, ME). Compound genotypes were generated by intercrossing the various cre genotype.

AAV^FLEX-hM3Dq, AAV^FLEX-hM4Di, AAV^{FLEX-TetTox-GFP}, and control viruses (e.g. AAV^GFP and AAV^cre-GFP) (Sternson and Roth, 2014; Carter et al., 2015; Armbruster et al., 2007) were prepared by the University of Michigan Viral Vector Core. The AAVs used in the manuscript were all serotype AAV8.

For injection, following the induction of isoflurane anesthesia and placement in a stereotaxic frame, the skulls of adult mice were exposed. After the reference was determined, a guide cannula with a pipette injector was lowered into the injection coordinates (NTS: A/P, –0.2; M/L,±0.2; D/V, –0.2 from the obex) and 100 nL of virus was injected for each site using a picospritzer at a rate of 5–30 nL/min with pulses. Five minutes following injection, to allow for adequate dispersal and absorption of the virus, the injector was removed from the animal; the incision site was closed and glued. The mice received prophylactic analgesics before and after surgery. The mice injected with AAV^Flex-hM3Dq, AAV^Flex-hM4Di, AAV^{Flex-TetTox-GFP}, or control viruses were allowed at least 1 week to recover from surgery before experimentation. AAV^FLEX-hM3Dq was injected into mice as a control for mice injected with AAV^FLEX-TetTox.

Lepr^TetTox, Cck^TetTox, LC^TetTox, and LCK^TetTox mice and their controls were monitored from the time of surgery for chow feeding. For stimulation studies, DREADD-expressing mice and their controls that were at least 3 weeks post-surgery were treated with saline or drug (CNO, 4936, Tocris) at the onset of dark cycle, and food intake was monitored over 4 hr. For chronic food intake and body weight changes in DREADD-expressing animals, mice were given saline for 2–3 days prior to injecting saline or drugs twice per day (approximately 5:30 PM and 8:00 AM) for 2 or 3 days, followed by saline injections for another 1 or 3 days to assess recovery from the treatment.

Mice were anesthetized with a lethal dose of pentobarbital and transcardially perfused with phosphate-buffered saline (PBS) followed by 10% buffered formalin. Brains were removed, placed in 10% buffered formalin overnight, and dehydrated in 30% sucrose for 1 week. With use of a freezing microtome (Leica, Buffalo Grove, IL), brains were cut into 30 μm sections. Immunofluorescent staining was performed using primary antibodies (FOS, #2250, Cell Signaling Technology, 1:1000; GFP, GFP1020, Aves Laboratories, 1:1000; dsRed, 632496, Takara, 1:1000), antibodies were reacted with species-specific Alexa Fluor-488,–568 or –647 conjugated secondary antibodies (Invitrogen, Thermo Fisher, 1:200). Images were collected on an Olympus (Center Valley, PA) BX53F microscope. Images were pseudocolored using Photoshop software (Adobe) or Image J (NIH).

Two types of CTA assay were conducted, as described previously (Cheng et al., 2020a): Saccharin CTA: Mice that were 8 weeks of age or older were individually housed in standard cages with low wire tops and free access to food and the lixits were removed. Mice were habituated to two water-containing bottles for 3–5 days until they learned to concentrate their daily water consumption into these two water bottles. On the conditioning day, the mice received only two saccharin (0.15%, 240931, Sigma) bottles. Following the 60 min exposure to saccharin, mice were injected intraperitoneally with the desired stimulus (vehicle control (0.9% NaCl), lithium chloride (0.3 M, 203637, Sigma)) in a volume equivalent to 1% of each animal’s body weight (10 mL/g), or CNO (1 mg/kg, 4936, Tocris Bioscience) Access to the two saccharin bottles continued for an additional 1 hr, followed by the return of normal water bottles. Two days later, each mouse received access to two water bottles (one containing 0.15% saccharine, the other containing water), and the amount of fluid ingested from each water bottle was measured.

HFD CTA: following an overnight fast, mice were provided with one hour access of HFD (D012492, Research Diets), followed with the relevant stimuli, followed by an extra one hour of HFD access before returning of normal chow. On the post-conditioning day, fasted mice received access to both HFD and chow and the consumption of each was measured.

Mice had ad libitum access to water and 60% lard-based HFD (Research Diet, New Brunswick, NJ; Cat. No. D12492) for 6–8 weeks prior to undergoing Sham or VSG surgery. Under isoflurane/O₂ mixture anesthesia, all mice received a midline incision in the ventral abdominal wall and the stomach was exposed. For VSG, approximately 80% of the stomach was transected along the greater curvature using an Echelon Flex Powered Vascular 35 mm Stapler Model (PVE35A; Ethicon endo-surgery) creating a gastric sleeve. The Sham surgery was performed by the application of gentle pressure on the stomach with blunt forceps for 15  s. All mice received one dose of Buprinex (0.1  mg/kg) and Meloxicam (0.5  mg/kg) immediately after surgery. Post-operatively, all mice received 1 ml warm saline subcutaneously on the first postoperative day and analgesic treatment with meloxicam (0.5 mg/kg) for 3 consecutive days. Animals were placed on DietGel Boost (ClearH₂O; Postland ME) for 2 days before and 3 days after surgery before pre-operative solid diet (60% HFD) was returned on day 4. Body weight and general health were observed daily for the first 10 days post-surgery. Food intake and body weight were measured weekly for the duration of the study. Body composition was measured using nuclear magnetic resonance (EchoMRI–900, EchMRI LLC, Houston, USA) before surgery and periodically throughout the study.

During glucose tolerance tests, all mice were fasted for 4–5 hr prior to oral administration of dextrose (2 g/kg body weight). Tail vein blood glucose levels were measured using Accu-Chek glucometers (Accu-Chek Aviva Plus, Roche Diagnostics) or Biosen C-line glucose analyzer (EKF diagnostic) at 0, 15, 30, 45, 60, 90, and 120 min post glucose administration. Gastric emptying rate was assessed by an oral gavage of glucose (2 g/kg) with acetaminophen (100 mg/kg) in 4–5 hr fasted mice. Blood was collected from the tail vein at baseline and 10  min after gavage in EDTA-coated microtubes. Plasma acetaminophen levels were used to assess the rate of gastric emptying and were measured using spectrophotometry assay (Sekisui Diagnostics). Plasma insulin levels were determined using ELISA colorimetric insulin assay kit (Crystal Chem).

A subset of mice was placed into an automated system equipped with metabolic chambers (TSE Systems International Group, Chesterfield, MO) to measure indices of energy homeostasis 5–7 week after surgery. Data were recorded continuously over the course of 7 days, and the last 4 days were used in the analysis.

Statistical analyses of the data were performed with Prism software (version 8). Two-way ANOVA, paired or unpaired t-tests were used as indicated in the text and figure legends. In detail, multiple groups with one variable comparison were performed using one-way ANOVA post hoc Tukey. Multiple groups with two variables comparison were performed using two-way ANOVA post hoc Sidak. Multiple group comparisons over time were performed using repeated measures ANOVA post hoc Tukey. All data are presented as Mean ± SEM. Two groups direct comparisons were performed using t-test. Two groups multiple comparisons were also performed using Multiple t-test with Bonferroni correction. Data were considered statistically significant when p<0.05.