Reduction of body weight by increased loading is associated with activation of norepinephrine neurones in the medial nucleus of the solitary tract

We previously provided evidence supporting the existence of a novel leptin‐independent body weight homeostat (“the gravitostat”) that senses body weight and then initiates a homeostatic feed‐back regulation of body weight. We, herein, hypothesize that this feed‐back regulation involves a CNS mechanism. To identify populations of neurones of importance for the putative feed‐back signal induced by increased loading, high‐fat diet‐fed rats or mice were implanted intraperitoneally or subcutaneously with capsules weighing ∼15% (Load) or ∼2.5% (Control) of body weight. At 3–5 days after implantation, neuronal activation was assessed in different parts of the brain/brainstem by immunohistochemical detection of FosB. Implantation of weighted capsules, both subcutaneous and intraperitoneal, induced FosB in specific neurones in the medial nucleus of the solitary tract (mNTS), known to integrate information about the metabolic status of the body. These neurones also expressed tyrosine hydroxylase (TH) and dopamine‐beta‐hydroxylase (DbH), a pattern typical of norepinephrine neurones. In functional studies, we specifically ablated norepinephrine neurones in mNTS, which attenuated the feed‐back regulation of increased load on body weight and food intake. In conclusion, increased load appears to reduce body weight and food intake via activation of norepinephrine neurones in the mNTS.

WHO in 2022 estimated that more than 1.9 billion adults worldwide were overweight and that over 650 million of these were living with obesity.The number of persons with obesity is continuously increasing, contributing to a further decline in health in all age groups. 1,2 the available pharmacological treatments for obesity, glucagon-like peptide 1 (GLP-1) receptor agonists such as semaglutide, constitute the most effective currently available pharmacologic monotherapy.Studies have shown a weight loss of $12% from continued treatment over 6 months. 3Side effects of GLP-1 receptor agonists, among which nausea is the most common one, are mild in comparison with bariatric surgery. 4Combined GLP-1R and GIP-R agonists such as tirzepatide may be even more effective causing 25%-30% weight loss within 1-2 years. 5 develop new potential obesity treatments, it is crucial to identify pathways involved in the homeostatic regulation of body weight.
7][8][9] Leptin is produced by white adipose tissue and then released into the bloodstream.It is known to suppress fat mass by targeting key hubs in the energy balance circuitry, such as the arcuate nucleus (ARC) of the hypothalamus.There was a general expectation that treatment with exogenous leptin receptor agonists would suppress fat mass.However, endogenous leptin levels in blood are increased in obese individuals.Moreover, pharmacological targeting of leptin receptors appeared to have little effect on common obesity.[12] Recently, we provided evidence for a second homeostatic system for the regulation of body weight besides the leptin system.We named this novel system the "gravitostat."The gravitostat hypothesis postulates that body weight is registered by sensors in osteocytes of weight carrying bones in the lower extremities. 13Activation of the gravitostat has previously been shown to decrease food intake and body weight. 13,14In this paper, we hypothesize that a brain mechanism might be involved in this homeostatic regulation by the gravitostat.However, it is completely unknown which brain regions are involved in receiving and processing gravitostat information, a gap in knowledge that we seek to address.
One brain area of recent interest for its involvement in regulating food intake and body weight is the nucleus of the solitary tract (NTS) of the brainstem.For instance, the medial nucleus of the solitary tract (mNTS) has been shown to receive and process information about the energy status of the body. 15,16In addition, the NTS is an important relay station for afferent information to the hypothalamus. 172][23] For example, the strongly anorectic drugs amphetamine and cocaine increase central norepinephrine levels and thereby decrease hunger and body weight. 24Pharmacological treatments influencing the central norepinephrine system, as part of combination drugs, have also been on the market for treatment of obesity but withdrawn due to side effects from the other component. 25e overall aim of our study was to identify downstream mediators of the gravitostat effect in the brain.To this end, we sought to identify population(s) of brain cells involved in gravitostat signaling, that (1) are activated by increased Load and (2) impact on food intake and body weight.We explored expression patterns of FosB (and, in some studies Fos) in Increased Load and Control animals.Based on location and function, a candidate population emerged that we identified neurochemically and also functionally by exploring the increased Load phenotype (food intake and body weight) when these cells were ablated.

| Animals
Male C57/B6J mice and Sprague-Dawley rats (mice 12-14 weeks and rats 20 weeks old at time of Load surgery) were purchased from Taconic Biosciences (Borup, Denmark).Animals were then put on 60% kcal fat (lard) high-fat diet (HFD, D12892; Research Diets, New Brunswick, NY, USA).Animals had free access to water and ad libitum HFD, and were kept under standardized conditions (12 h light/dark cycle [lights on at 7:00 a.m.], 50%-60% humidity, 21 C temperature).The local ethics committee for animal care at the University of Gothenburg approved all animal procedures, ethical approval number 3308/20.manufactured in-house) under isoflurane anesthesia at 12-20 weeks of age.Prior to surgery, capsules were either filled with wolfram granulate (Edstraco AB, Rönninge, Sweden) weighing $15% of the body weight (load) or left empty, weighing $2.5% of the body weight (control).Capsules were attached with suture to the peritoneum or unsutured on the lower back, after which the peritoneum was closed with suture and the skin was closed with agraff.Alternatively, capsules were implanted s.c. at the lower back, after which the skin was closed with agraff.An analgesic (5 mg mL À1 Rimadyl ® , dose: 1 mL kg À1 mouse; Orion Pharma Animal Health, Solletuna, Sweden) was injected s.c.Rats and mice were individually housed after surgery, and sacrificed on Days 3-14 depending on the experiment.

| Food intake measurements
Mice were single-housed and had access to pre-weighed HFD on the cage floor, $16-18 g.This food was carefully collected and weighed every day at 7:00 a.m. when lights turned on in the animal room, and refilled as needed.

| Saporin toxin microinjections
To lesion norepinephrine neurones in the medial part of the NTS, anti-dopamine-B-hydroxylase (DBH) antibodies conjugated to the neurotoxin saporin (Advanced Targeting Systems, Carlsbad, CA, USA) or vehicle was injected bilaterally into the NTS.Anti-DBH conjugated saporin enters neurones containing the catecholamine-synthesizing enzyme DBH, thereby ablating them. 26,27Injection sites encompassed the NTS region that contains the highest incidence of colocalization between FosB-and DbH-positive neurones after increased Load treatment, corresponding to the location of the A2 cell group. 19,28For analgesis, 500 μL 1% karprofen (5 mg mL À1 Rimadyl ® , dose: 1 mL kg À1 mouse; Orion Pharma Animal Health, Solletuna, Sweden) was injected.Mice were anesthetized with Sedastart (Omnidea, Stockholm, Sweden) and secured in a stereotaxic frame using blunt ear bars, with the head horizontal.The skin over the dorsal neck surface was shaved and incised, and the neck muscles were retracted and bluntly dissected to expose the meninges overlying the dorsal surface of the caudal medulla.The meninges were gently split and the skull exposed.A 10 μL Hamilton syringe filled with toxin or vehicle was positioned on the midline at bregma and then moved À7.5 mm area postrema (AP) and ± 0.3 mm DV.A hole was drilled and the Hamilton syringe needle was lowered 4.8 mm.At each injection site, 1 μL of sterile 0.15 M NaCl vehicle containing 0 or 5 ng of toxin was delivered by an injection pump over 2 min.The toxin was freshly prepared from a frozen stock solution within 2 h of injection.
After surgery, mice were injected with Sedastop (Omnidea, Stockholm, Sweden) and allowed to recover on a heating pad.Mice were then transferred to their home cages where they were allowed to recover for 2 weeks with ad libitum access to HFD.
As a control for the secondary antibodies, some sections were incubated with mismatching primary and secondary antibodies, resulting in negative staining (as a control for unwanted cross-reactivity).
Primary antibodies, their dilutions, catalogue numbers and resource indentification incentive number (RRID), as well as the manufacturers providing them are listed in Table 1.

| Confocal microscopy and cell counting
Fluorescent images of the stained sections were obtained using either a confocal microscope system (LSM 700; Zeiss, Oberkochen, Germany), together with a Plan APO Â 40 A/1.40 oil lens (for higher magnification pictures) or a Plan Fluor Â20/0.75 lens (for anatomical overview pictures) and a solid-state laser.For co-localization, focus stacking was used for magnified images to achieve a greater depth of field and as such make it possible to more accurately detect possible co-localization.
Triple channel confocal images (to cover the entire NTS) were generated with a Plan Fluor Â20/0.75 lens and a solid-state laser.A tile scan of 3 Â 3 tiles was obtained from the center of the NTS, covering the entire nucleus.The emission spectrum of the secondary fluorescent antibody is well known.By adjusting the splits of the confocal microscope, the signal of the fluorophore was maximized while minimizing background fluorescence.The evaluation of cell labeling was performed with the cell nucleus in the plane of focus.Grayscale images of DAB-staining were obtained using a Leica DMRB microscope and a Leica DFC7000T camera (Leica Microsystems, Wetzlar, Germany).
For each counted region, the number of cells for one animal were blindly counted unilaterally from the same side on each section and divided by the number of sections counted.An average of each region was then obtained per treatment group by adding the average cells per animal together and then dividing by the number of animals in the group.Regions were defined using an anatomical atlas. 29,30 determine successful depletion of NE-neurons in the mNTS, tyrosine hydroxylase (TH)-cells were blindly counted bilaterally.
Saporin-treated mice that had more than 30% of TH-positive neurons intact were counted as misses and excluded from the study.

Counting was performed without knowledge of what treatment
the animal had received.Micrographs were adjusted for brightness and contrast in FIJI 31 and tile scan overviews were constructed using a FIJI plug-in. 32

| RNA-scope
For RNAscope, NTS-containing brain sections (12 μm thick) were cut and fixed in 10% formalin (Thermofisher scientific, Waltham, MA) for 30 min.Following two quick washes in 1Â PBS, brain slices were dehydrated in 50% (5 min), 70% (5 min), and 2 Â 100% (5 min each) ethanol and treated with protease solution (pretreatment IV) at room temperature for 30 min.The protease was washed away with PBS.Target probes (see Table 1) and negative control probes were applied directly on the sections to cover them completely and incubated at 40 C for 120 min in the HybEZ oven.Next, we applied preamplifier and amplifier probes (AMP1, 40 C for 30 min; AMP2, 40 C for 15 min; AMP3, 40 C for 30 min).Finally, brain sections were incubated for 0.5 min with DAPI and mounting medium for fluorescence (ThermoScientific, Waltham, MA, USA).Fluorescent images of the NTS were captured at 20Â (overview) or 40Â (magnifications) as z-stacks using LSM700 Zeiss confocal microscope.These z-stacks were later processed with Zen software into maximum intensity projections, where the individual pictures from the z-stack were superimposed into one image.

| Increased Load activates neurones in mNTS of the brainstem
We hypothesized that the gravitostat, a homeostatic regulator of body weight, regulates food intake and body fat mass by activating neurones in parts of the brain influencing energy balance in general.
To test this hypothesis, some HFD-fed rats were implanted with empty capsules (Control; $2.5% of total body weight; Figure 1A), while others received capsules with an increased Load (Load; $15% of body weight; Figure 1B).NTS micrographs indicated that, at 4 days after capsule implantation, Increased Load increased the number of FosB immunoreactive cells in rat (Figure 1B) and Fos immunoreactive cells in rat (Figure S1), all relative to cell counting in Control rats (Figure 1A).Indeed, at this time point, there was a two-fold increase in the number of cells detected that express FosB in rat mNTS ( p = .003;Figure 1C).Increased Load rats lost more biological bodyweight than Control animals.The effect of treatment was statistically significant but not the interaction of treatment over time (treatment: p = .034,treatment over time: p = .285).By the day of sacrifice, Increased Load rats (Figure 1D) had lost $4% more biological body weight than Control rats.
In another part of the dorsal vagal complex, the AP, we did not find a statistically significant increase in FosB immunoreactivity in rodents with increased i.p.Load (Figure S2A,B, p = .12[mouse], p = .21[rat]).
We also investigated potential differences in FosB immunoreactivity in the lateral parabrachial nucleus (lPBN).Here we also did not find a statistically significant difference between Load and Control mice or rats (Figure S2C-H, p = .58[mouse], p = .62[rat]).
In the ARC of the hypothalamus (Figure S3), there was an increase in the number of FosB-containing cells from mice with of Increased Load (Figure S3D) compared to Control (Figure S3B).The results of cell counting confirmed this observation Figure (S3E).There was also an increase in Neuropeptide-y (NPY) containing cells in Increased Load (Figure S3D) mice compared to Control (Figure S3B) mice.The latter finding is in line with the earlier report that Increased Load leads to enhanced NPY mRNA in the ARC.Cumulative food intake over 5 days showed a marked difference between the vehicle-treated Load and Control groups (p ≤ .001)but not in the saporin-treated mice (p ≤ .446)(Figure 5H).
Immunohistochemical mNTS staining from vehicle-treated Control and Load mice showed a significant difference in FosBimmunoreactive cells ( p = .002;Figure 5I,J,M).However, no such difference was found in saporin-treated Control and Load mice ( p = .909;Figure 5K,L,N).
Intact norepinephrine synthesizing neurones in the mNTS thus seem to be necessary for the body weight and food intake reducing effects of increased Load.

| DISCUSSION
We have hypothesized the existence of a new homeostat that uses gravity to keep body weight constant and coined it the "gravitostat." Activation of the gravitostat is seen after implantation of an increased Load.We herein confirm earlier results that carrying an increased Load, by implantation of a heavy capsule, decreases body weight and food intake. 13,14Given that HFD-fed animals bearing an increased Load reduced their food intake, a brain mechanism has been proposed.In our study design, we were interested in studying potential activation of brain nuclei of importance for body weight homeostasis at the earliest time point where we saw a difference in body weight.This was assessed using Student's T test.However, when we later analyzed the body weight data from i.p.Load rats and s.c.Load mice using repeated measures ANOVA, the interaction effect of treatment and time did not end up being significant.However, we still saw a significant effect on treatment alone.This is likely to be due to the study design, where we were focused primarily on the mechanistic effects on the brain and not primarily on body weight differences.To identify increased Load activated neurones in the brain, we exploited classic functional mapping techniques.Acutely activated neurones were identified by immunohistochemistry for the transcription factor c-fos (that gives rise to Fos protein 33 ) as well as for FosB, a related factor that is regulated more slowly than Fos.FosB can be maintained for up to 12-24 h, 34 enabling measurements for longer time after the surgical trauma of increased Load implantation.While an acute induction of Fos and FosB immunoreactivity is likely to be caused by the surgical trauma induced by capsule implantation, leaving the animals until a clear difference in loss of biological body weight is visible, is likely to minimize this risk.
Another factor that might influence body weight and Fos/FosB immunoreactivity is stress and/or nausea.We have previously measured locomotor activity, peripheral corticosterone and norepinephrine levels, as well as inflammatory cytokines, and seen no differences between Load and Control animals. 14,35However, we still cannot completely exclude the possibility that the effect of load, at least partly, is due to interaction with non-specific factors such as stress and inflammation.Moreover, the differences in effects between subcutaneous and intraperitoneal load, could be due to differences in non-specific influence.However, while an acute induction of Fos and FosB immunoreactivity is likely to be caused by the surgical trauma of capsule implantation, leaving the animals until a clear difference in loss of biological body weight is visible, is likely to minimize the interference by trauma, leaving the gravitostat effect.
Functional mapping studies were performed in several brain nuclei known to be of importance for body weight regulation.AP, NTS, lPBN, central amygdala (CeA), ARC, paraventricular nucleus, and lateral hypothalamic area were investigated.Our studies revealed that while there was general activity in all of these nuclei in both Load and Control animals, only in the ARC and the mNTS was there a statistically significant difference between the two groups.We have previously seen an increase in NPY/AgRp mRNA in the ARC of Load mice, so one possible way to interpret the increased activation in this nucleus might be as a failed compensation mechanism. 13We also performed cell counting on AP and lPBN.For the AP, there was a slight tendency for FosB to be increased in Load mice, as compared with Control mice, but this was not statistically significant.In rat AP and both rat and mouse lPBN, no such differences were observed.
Previously, we showed that osteocytes in weight bearing limbs sense changes in body weight. 13We assumed that this information is transmitted via either hormonal or neuronal signals to the brain, to affect food intake and other behaviors, but there has been very little data on this.Our Fos and FosB mapping studies identify the mNTS, in the neural response to carrying an increased Load.We saw an effect on mNTS-activation both with intraperitoneal and subcutaneous Load.While the effect on body weight of subcutaneous Load is smaller than intraperitoneal Load, this indicates that the entirety of the effect is unlikely to be mediated solely by the gut-brain axis.More specifically, the mNTS is of importance for integrating post-prandial gut signals and promoting meal termination. 15,17,36,37In addition, it has previously been suggested that the mNTS is an integrating center for combining information about energy availability from the gut via afferent vagus and from adipose tissue via leptin. 17However, it is well-known that the AP is a target for signals of malaise and stress from the periphery, and the same has also been proposed to be true for the mNTS. 38We here propose a third afferent signaling pathway influencing mNTS.The gravitostat, a regulator of body weight homeostasis, could provide information about body weight, as measured by osteocytes in lower extremities.These osteocytes could then induce a neuronal or endocrine signal, reaching mNTS, eventually resulting in body weight reduction and decreased food intake.
Roughly 55% of activated neurones of mNTS contained TH and DbH, indicating that they produce norepinephrine, as also supported by their anatomic localization in the mNTS.However, this still leaves roughly 45% of activated neurons that do not express these enzymes.
0][41] It remains to be investigated if these substances could also play a role in the gravitostat system.Based on this information, we believe these activated neurones to be mainly norepinephrinergic.This assumption was supported by the fact that many of the Load (FosB)activated neurones were located in the mNTS, a part of the A2 area, a well-known site of norepinephrine production as described by Hillarp and colleagues. 18,20Norepinephrinergic mNTS neurones are known to project to brain areas important for body weight regulation; the lPBN, ARC, and CeA, via the ventral noradrenergic bundle (VNAB). 42,43 here show that chemical ablation of mNTS neurones containing DbH (i.e., norepinephrine) markedly attenuates the reduction of body weight and food intake seen in HFD-fed mice with increased Load.Intact norepinephrinergic NTS neurones thus appear to be necessary for these effects of increased Loading.However, it is still unclear if norepinephrine is the important signaling substance in these cells, or if other neuropeptides also expressed therein mediate the effect.
It has previously been shown that norepinephrinergic neurones are involved in regulation of body weight and food intake in general. 22,40ch of the information comes from the pharmacological studies.The anorexic effects of amphetamine, for example, is known to be caused by increased release of norepinephrine. 24Certain other pharmaceuticals, such as the amphetamine-type drug phentermine, are also known to act via the norepinephrine system. 24Phentermine, in combination other drugs such as fenfluramine and topiramate, was used as treatment for obesity but withdrawn mainly due to side effects. 25Serotonin and norepinephrine reuptake inhibitors (SNRIs) have also been shown to induce satiety. 44Conversely, blocking norepinephrinergic signaling via ablation of the VNAB leads to overeating. 23Chemical ablation of NTS norepinephrine neurones also attenuates the anorectic response to pharmacologic cholecystokinin treatment. 28e limitation of this study is that we do not see a statistically significant interaction effect on body weight for treatment over time in the s.c.mouse and i.p. rat experiments.A possible reason for this lack of effect could be that we have few measured timepoints, and that the experiment continued only for 3-4 days.Rodents in these short-term experiments were sacrificed as soon as we saw a significant body weight difference between Load and Control using a Student's t-test, for the purpose of investigate differences in neuronal activation when the effect of Load was first observed.After 3-4 days, the Load and Control groups often continue to deviate from each other in body weight, thereby increasing the chances of getting significant interaction effects.We have previously seen a statistically significant effect on body weight in longer term s.c. and i.p. experiments. 13,14rther limitations of this study include the possibility that the effect of Load on body weight and food intake seen here is mediated by a gut-brain axis, and not the Gravitostat.We cannot in this study full exclude this possibility, although our finding that subcutaneous implantation of capsules also leads to a reduction in body weight, in both this and previous experiments, suggests that the entire reduction is unlikely to be mediated solely via the gut-brain axis. 13However, it should be acknowledged that the body weight suppression by intraperitoneal load is larger than the suppression of body weight by subcutaneous load.Moreover, we have previously found that the suppression of body weight and fat mass by increased Load is dependent on osteocytes. 13Thus, several lines of evidence suggest that there is a bone-brain axis, as well as a gut brain axis, regulating food intake and body weight.
A perhaps surprising finding is that NE-ablation by DbHconjugated saporin did not affect the amount of weight gained by mice after surgery.This has previously been shown to be the case in rats, where saporin-treated rats gain more body weight than vehicletreated ones. 27,28This might be due to species differences, as to our knowledge DbH-conjugated saporin has not been used in the mouse brain before.Our interpretation of this data is that the hypophagic effects of Load are attenuated in NE-ablated mice.This is a likely explanation for the loss of body weight reduction seen in this group.
An alternative interpretation of the data is that NE-ablation gives rise to hyperphagia in mice as well, but that in the absence of a stressor, such as Load surgery, this effect is hidden.
There is little doubt that endogenous leptin plays a very important role in preventing severe obesity. 106][47] We have hypothesized that the gravitostat could be such a factor. 14 summary, we hereby propose that there exists a population of neurones in mNTS of the brainstem that are stimulated by gravitostat activation as indicated by FosB/Fos expression in response to increased Load.Many of these gravitostat-activated mNTS neurones appear to contain norepinephrine.These results suggest that many of the gravitostat-activated neurones produce norepinephrine.Norepinephrine, as well as many known pharmacologic modulators of the norepinephrine system, are well-known suppressors of food intake and body fat. 48In this study, we found that ablation of norepinephrine-producing cells in the mNTS attenuated the decrease in body weight and food intake induced by increased Load in HFDfed mice.Thus, the present results suggest a novel physiological role for norepinephrine in the mNTS, as being part of the gravitostat, a homeostatic regulator of body weight.However, other interpretations of our data are also possible, as mentioned above.More studies are needed to confirm our particular interpretation.Further research is also needed to identify additional signaling substances to norepinephrine, including neuropeptides, that may participate in the gravitostat associated regulation of body weight, fat mass, and food intake.
Moreover, the neurocircuits and target cells of the gravitostatactivated norepinephrine cells should be investigated.writingreview and editing.
Rats and mice were divided weight-matched Control and Load groups (see Figure S6 for average group weights at time of load surgery).Rats and mice were i.p. or s.c.implanted with capsules (60 Â 10 Â 10 mm [rat], 25 Â 8 Â 10 mm [mouse i.p.] and 18 Â 8 Â 8 mm [mouse s.c.], Repeated measures ANOVAs for biological body weight and food intake changes over time were performed in IBM SPSS Statistics version 29.0.0.0.If Mauchly's test of sphericity was significant (p ≤ .05)for treatment Â time, corresponding values were reported.Otherwise, values under Greenhouse-Geisser were reported.Data are presented in Section 3 for effects of treatment and interaction between treatment and time.For a full outcome of the statistical tests for effect of treatment and treatment Â time (Sum of Squares [SS], df [degrees of freedom], mean square [MS], F, p-value, and error margin), see Figure S7.Two-tailed, unpaired, Student's t-tests with equal variance assumed were performed in Microsoft Excel, for cell counting data and cumulative food intake.Graphs were made using GraphPad Prism 9 version 9.4.1.

3. 3 |
Figure 3E-H).Cell counting shows $55% of FosB immunoreactive neurones in mNTS of Load mice are also TH or DBH immunoreactive, whereas co-localization is found in only $15% of FosB immunoreactive neurones in mNTS of Control mice (Figure 3I, p ≤ .001).Thus, we could conclude that Load significantly increases the activation of THand DbH neurones in the mNTS.Using RNA-Scope, we confirmed this co-localization between FosB and enzymes vital for the synthesis of catecholamines (Figure 4A-D).Cell counting shows a similar pattern on the mRNA as

3. 4 |
Pharmacological ablation of norepinephrine synthesizing neurones in the NTS attenuates the suppression of food intake and body weight caused by increased Load Pharmacological ablation of norepinephrine-synthesizing neurones in the mNTS was achieved by local injections of anti-DbH conjugated saporin in HFD-fed mice.This substance is internalized in DbH-producing cells and irreversibly inhibits protein synthesis, which results in cell death.Mice were allowed to recover for 2-3 weeks and their body weights recorded (FigureS5).Load surgery was performed according to Section 2.At time of sacrifice, DbH-immunohistochemistry was markedly reduced in the mNTS in mice injected with anti-DbH conjugated saporin, as compared with vehicle injected mice (Figure5A,B; FigureS5A,B).Vehicle injected increased Load mice lose $9% more biological body weight by Day 14, as compared with vehicle-injected Control mice ( p = .005)(Figure5C).In contrast, anti-DbH conjugated saporin injection to increased Load mice were statistically similar to their Control mice ( p = .829)(Figure5D).A short-term follow-up experiment was performed where HFDfed mice were divided into groups as mentioned above, and were injected either with saporin or vehicle.After 2 weeks of recovery, Load surgery was performed according to Section 2. Mice were sacrificed 5 days after Load surgery, and at this time, vehicle-injected Load mice had lost $6% more biological body weight than Control mice ( p ≤ .001).In the saporin-treated group, Load mice were statistically similar to Control mice ( p = .255)(Figure 5E,F); 24-h food intake was measured every day at onset of light phase.Saporin-treated Load mice were more similar in food intake to their Control littermates than the vehicle-treated groups, but there was no statistically significant change in treatment over time ( p = .166[vehicle]; p = .780[saporin]) (Figure 5G).However, for treatment alone, there was a statistically significant decrease in vehicle-treated Load mice, as compared with Control mice ( p = .001).

F I G U R E 2
Increased Load implanted subcutaneously activates neurons in the medial nucleus of the solitary tract (mNTS) in mice.(A,B) FosB immunoreactivity in the mNTS in mice sacrificed 3 days after subcutaneous Load surgery (n = 9 per group).Load brains showed a marked increase in FosB immunoreactivity as compared with Control brains.(C) Cell counting in the mNTS verifies the increase in FosB-immunoreactive cells in brains from Load mice (Student's t-test, p = .004).(D) Subcutaneous increased Load causes increased biological body weight reduction as compared with subcutaneous Control when investigated for treatment alone (Repeated Measures ANOVA, p = .032)but not for a potential interaction between treatment and time (Repeated Measures ANOVA, p = .132).AP, area postrema; CC, central canal; NTS, nucleus of the solitary tract; Bregma À7.32 to À7.76 mm; Scale bars = 100 μm.**p < .005,#p < .05.

F I G U R E 3
Neurons activated by increased load in the medial nucleus of the solitary tract (mNTS) in mice express markers of synthesis of norepinephrine at the protein level.Fluorescent immunohistochemistry was performed on Load and Control brain sections containing mNTS taken from mice sacrificed 5 days after intraperitoneal capsule surgery (n = 6 in each group; representative images shown).(A-D) Immunohistochemical staining for FosB immunoreactive (red) and tyrosine hydroxylase (TH)-containing (green) neurons in Control (A,B) and Load (C,D) mice.Micrographs (B) and (D) are magnifications of the areas indicated in (A) and (C), with each channel shown separately.(E-H) Immunohistochemical staining for FosB-(red) and dopamine beta hydroxylase (DBH)-(green) immunoreactive neurons in Control (E,F) and Load (G,H) mice.Micrographs (F) and (H) are magnifications of the areas indicated in (E) and (G) with each channel shown separately.(I) Cell counting shows an increase in co-localization between both TH and DBH with FosB in brains from Load mice, as compared with Control mice (Student's ttest p ≤ .001 in both comparisons).AP, area postrema; CC, central canal; NTS, nucleus of the solitary tract; Bregma À7.32 to À7.76 mm.Scale bars = 100 μm (overview images) and 20 μm (magnified images).***p < .001.

F I G U R E 4
Neurons activated by increased Load in the medial nucleus tractus solitarius (mNTS) of mice express markers of synthesis of norepinephrine at the mRNA level.RNAscope visualization of mRNAs for FosB, tyrosine hydroxylase (TH), dopamine-beta-hydroxylase (DbH) on sections containing containing the mNTS in Load and Control mice sacrificed at 5 days after intraperitoneal capsule surgery (n = 6; representative images shown).(A) Overview image of Control mouse NTS, showing FosB in red, TH in green and DAPI (nuclear stain) in blue.Subsequent images are magnifications of the area indicated in the overview, with each channel shown separately.(B) Overview image of Load mouse NTS, showing FosB in red, TH in green and DAPI (nuclear stain) in blue.Subsequent images are magnifications of the area indicated in the overview, with each channel shown separately.(C) Overview image of Control mouse NTS, showing FosB in red, DbH in green and DAPI (nuclear stain) in blue.Subsequent images are magnifications of the area indicated in the overview, with each channel shown separately.(D) Overview image of Load mouse NTS, showing FosB in red, DbH in green and DAPI (nuclear stain) in blue.Subsequent images are magnifications of the area indicated in the overview, with each channel shown separately.(E) Cell counting shows an increase in co-localization between both TH and DBH mRNA with FosB mRNA in brains from Load mice, as compared with Control mice (Student's t-test p = .006[TH], p = .027[DbH]).AP, area postrema; CC, central canal; Bregma À7.32 to À7.76 mm.Scale bars = 100 μm (overview images) and 20 μm (magnified images).*p < .05,**p < .005.F I G U R E 5 Legend on next page.ZLATKOVIC ET AL.

5
Ablation of dopamine-beta-hydroxylase (DbH) containing neurons in the medial nucleus of the solitary tract (mNTS) attenuates the suppression of biological body weight and food intake caused by increased load.(A,B): Immunohistochemistry of mNTS showing THimmunoreactivity (green) after bilateral injection into the mNTS of vehicle or DbH-conjugated Saporin toxin in mice.(C) Vehicle treatment does not affect reduction of biological body weight in increased Load mice as compared with Control mice (n = 8 per group, Repeated Measures ANOVA p = .005).(D) Saporin treatment attenuates the reduction of biological body weight in increased Load mice as compared with Control mice (n = 8 per group, Repeated Measures ANOVA p = .587).(E,F) A separate short-term experiment showed similar results, wherein vehicle treatment does not affect reduction of biological body weight in increased Load mice as compared with Control mice (n = 5-6 per group, Repeated Measures ANOVA p ≤ .001).Saporin treatment attenuates the reduction of biological body weight in increased Load mice as compared with Control mice (n = 5-6 per group, Repeated Measures ANOVA p = .255).(G) Increased Load does not significantly reduce per day food intake in vehicle treated, and saporin treated, mice (n = 5-6 per group, Repeated Measures ANOVA p = .166[vehicle], p = .780[saporin]).(H) Increased Load reduces cumulative food intake over five days in vehicle treated, but not saporin treated, mice (n = 5-6 per group, Student's ttest p ≤ .001(vehicle), p = .446(saporin)).(I-L) Immunohistochemistry shows increased FosB-immunoreactivity in the mNTS of vehicle-treated Load mice, as compared with vehicle treated Control mice, as well as saporin treated Load and Control mice.(M,N): Cell counting shows a significant increase in FosB-immunoreactive cells in vehicle Load mice, as compared with vehicle Control mice (Student's t-test p = .002).However, no such differences were found when comparing FosB-immunoreactive cells in saporin Load and Control mice (Student's t-test, p = .909).AP, area postrema; CC, central canal; NTS, nucleus of the solitary tract; Bregma À7.32 to À7.76 mm; Scale bars = 100 μm.**p < .005,***p < .001,#p < .05.

1
List of antibodies and probes used in this article.