Effects on synaptic control of dopamine within the NAc

Effects on the VTA

VTA dopaminergic output underlies motivated behaviours and is affected by consumption of a HFD ^{Reference Gantz, Ford, Morikawa and Williams24,Reference Butler and Eckel79,Reference Morales and Margolis80} . The VTA receives glutamatergic and GABAergic signals from local interneurons and limbic and cortical regions ^{Reference Morales and Margolis80} as well as local somatodendritic dopamine release that negatively regulates dopamine neuron excitability ^{Reference Gantz, Ford, Morikawa and Williams24} . Glutamatergic and cholinergic inputs activate receptors on VTA dopamine neurons to promote NAc dopamine release ^{Reference Wickham, Solecki, Rathbun, McIntosh and Addy81} . The VTA receives orexin from LH, NPY from orexigenic glucose-sensing ARC neurons ^{Reference Ferrario, Labouèbe and Liu2,Reference Gumbs, Vuuregge and Eggels82} , and anorectic POMC/CART signals from ARC ^{Reference Ferrario, Labouèbe and Liu2} . The VTA integrates this information related to homeostasis and environmental cues with dopamine neurons projecting to the NAc. However, HFD intake reduced VTA tyrosine hydroxylase (TH) mRNA ^{Reference Li, South, Han, Chen, Wang and Huang10} and protein ^{Reference Sharma and Fulton9,Reference Naneix, Tantot and Glangetas83} expression, which occurred regardless of obesity and was restored after switching to low-fat diet ^{Reference Li, South, Han, Chen, Wang and Huang10} . This showed HFD intake reduced capacity for dopamine synthesis, because TH is the rate-limiting enzyme in synthesising dopamine ^{Reference Levitt, Spector, Sjoerdsma and Udenfriend84} . Further, 6-week ad libitum HFD feeding attenuated D2R agonist quinpirole-induced inhibition of VTA dopamine neuron firing, suggesting D2R desensitisation ^{Reference Koyama, Mori and Kanamaru85} , whereas intermittent access to cafeteria diet for 3 weeks increased inhibitory effects of quinpirole and reduced ethanol and sucrose intake ^{Reference Cook, Hendrickson, Garwood, Toungate, Nania and Morikawa58} , showing sensitivity of VTA autoreceptor function to type and length of diet. HFD intake similarly reduced excitability of mouse VTA GABA neurons ^{Reference Koyama, Kawaharada and Terai86} that reduce NAc dopamine concentration ^{Reference Brown, Tan, O’Connor, Nikonenko, Muller and Lüscher69,Reference Smolders, De Klippel, Sarre, Ebinger and Michotte73} and stop behaviour (e.g. food intake) ^{Reference Brown, Tan, O’Connor, Nikonenko, Muller and Lüscher69–Reference Lee, Vogt, Rubenstein and Sohal71} . Overall, HFD and obesity affect VTA dopamine and GABA neuron protein expression and function, and a further diet-induced effect includes altered sensitivity to direct responses of VTA neurons to hormonal indicators of energy status.

VTA neurons express receptors for insulin and leptin ^{Reference Figlewicz, Evans, Murphy, Hoen and Baskin87} with leptin-receptor-expressing dopamine and GABA neurons projecting to the NAc ^{Reference Fulton, Pissios and Manchon88} , and activation of these receptors decreases food intake. Insulin in the VTA is important in reducing dopamine neuron activity to control behaviour, demonstrated by reduced locomotion and NAc dopamine after VTA application of insulin ^{Reference Naef, Seabrook, Hsiao, Li and Borgland89} . In the VTA, insulin reduced somatodendritic dopamine release and hedonic feeding, which was abolished by blocking the DAT ^{Reference Mebel, Wong, Dong and Borgland90} . Furthermore, insulin induced AMPAR- and endocannabinoid-mediated long-term depression in VTA dopamine neurons which was attenuated by genetic or diet-induced elevation of insulin ^{Reference Labouèbe, Liu and Dias91,Reference Liu, Labouèbe, Karunakaran, Clee and Borgland92} , showing a role of insulin in the VTA to signal satiety that is inhibited by HFD. VTA signalling is also affected by leptin resistance, as HFD consumption reduced ability of VTA leptin administration to limit food intake and weight gain in obesity-prone rats ^{Reference Bruijnzeel, Qi and Corrie93} and induced leptin resistance specifically in the VTA and ARC ^{Reference Matheny, Shapiro, Tümer and Scarpace41} . Diet-induced obesity also spurred leptin resistance in ARC NPY/AGRP and POMC/CART neurons ^{Reference Enriori, Evans and Sinnayah42} and LH neurotensin–galanin–GABA neurons ^{Reference Opland, Sutton and Woodworth94,Reference Leidmaa, Gazea and Patchev95} necessary for reducing food intake by inhibiting orexin and activating VTA neurons ^{Reference Ferrario, Labouèbe and Liu2,Reference Opland, Sutton and Woodworth94,Reference Leidmaa, Gazea and Patchev95} . Therefore, lack of hypothalamic and VTA insulin and leptin signalling may reduce capacity for NAc dopamine release to promote compensatory over-seeking of palatable food.

Effects on anorectic hormones in the NAc

Insulin gains access to the brain via transport across the blood–brain barrier ^{Reference Zeng, Zhang and Hu96} and local production in the brain ^{Reference Mehran, Templeman and Brigidi97} , and activation of tyrosine kinase receptors by insulin promotes phosphatidylinositol-3 kinase (PI3K) and protein kinase B (Akt) or the mitogen-activated protein kinase (MAPK)/extracellular-signal-regulated kinase (ERK) pathways, the latter which is also known as the Ras-Raf-MAPK/ERK kinase (MEK)-ERK cascade ^{Reference Zeng, Zhang and Hu96} . Insulin valuates food reward by fine-tuning dopamine neurotransmission at NAc dopamine terminals. For example, NAc core and shell insulin administration increased dopamine release and reuptake in control animals but not in animals consuming HFD ^{Reference Fordahl and Jones34,Reference Stouffer, Woods and Patel40} . Furthermore, HFD-induced impairments were reversed by promoting insulin receptor substrates, while effects of insulin were abolished by inhibiting insulin receptor or PI3K ^{Reference Fordahl and Jones34} , implicating HFD-induced insulin resistance in impaired dopamine control. Insulin receptors expressed on NAc dopamine neuron terminals ^{Reference Stouffer, Woods and Patel40} activate Akt and ERK to shuttle DAT to the plasma membrane to promote dopamine reuptake ^{Reference Nash98} . Dopamine release is also promoted by insulin as NAc CINs expressed insulin receptors at high density ^{Reference Stouffer, Woods and Patel40} and released acetylcholine in response to insulin ^{Reference Stouffer, Woods and Patel40,Reference Patel, Stouffer, Mancini, Nicholson, Carr and Rice43} , supporting necessity of insulin in encoding sucrose preference ^{Reference Stouffer, Woods and Patel40} . While an obesogenic diet blunted insulin-induced NAc dopamine release and reuptake ^{Reference Patel, Stouffer, Mancini, Nicholson, Carr and Rice43} , food restriction alternately enhanced insulin receptor expression and stimulated dopamine reuptake ^{Reference Jones, Woods, Zhen, Antonio, Carr and Reith50} . Similar effects have been shown with impaired leptin signalling, as leptin-deficient mice had reduced electrically evoked NAc shell dopamine release and reduced TH and DAT expression ^{Reference Fulton, Pissios and Manchon88} , whereas leptin increased activity of NAc DAT and TH and increased amphetamine-evoked dopamine release ^{Reference Perry, Leinninger and Chen99} . Leptin activates intracellular signalling cascades similarly to insulin in addition to the Janus-activated kinases (JAK)–signal transducers and activators of transcription (STAT)–suppressors of cytokine signalling (SOCS) pathway that reduce NPY/AGRP but increase POMC/CART and LH neurotensin neuronal activity ^{Reference Ferrario, Labouèbe and Liu2,Reference Houseknecht and Spurlock100} . Further, leptin is also transported into the brain ^{Reference Banks, Kastin, Huang, Jaspan and Maness101} and is expressed in plasma and cerebrospinal fluid in proportion to adipocyte size ^{Reference Hamilton, Paglia, Kwan and Deitel102} and adiposity ^{Reference Schwartz, Peskind, Raskind, Boyko and Porte103} , suggesting that leptin conveys energy sufficiency. However, individuals with obesity have elevated plasma leptin ^{Reference Houseknecht and Spurlock100} , supporting leptin resistance as a comorbidity of obesity. Collectively, these studies showed NAc insulin promotes reward seeking by encoding reward salience via increased dopamine release and maintenance of dopamine reuptake, and, whereas food restriction may prime dopamine responses through insulin, HFD-induced insulin and leptin resistance may reduce NAc dopamine neurotransmission to alternately promote food seeking. A putative contributor to leptin and insulin resistance associated with diet-induced obesity is chronic inflammation triggered by saturated fats and rapid adipose tissue expansion ^{Reference Wang, Liu and Wang38,Reference Lee, Wu and Fried104,Reference Tsukumo, Carvalho-Filho and Carvalheira105} .

Acute stress: role of HPA, CRF and inflammation on food intake

Acute stress encompasses a huge variety of physiological and psychological triggers that activate the hypothalamic–pituitary–adrenal (HPA) stress axis in a coordinated effort with metabolic, immune, autonomic nervous, and digestive systems to increase breathing and heart rate but slow digestion in preparation for ‘fight or flight’ response ^{Reference McEwen122,Reference Mcewen and Gianaros123} . Various stressors induce the hypothalamic paraventricular nucleus (PVN) to release corticotrophin-releasing factor (CRF), stimulating the pituitary gland to release adrenocorticotropin (ACTH) that targets adrenal glands to release glucocorticoids (corticosterone in rodents) and promote stress responses ^{Reference Rabasa and Dickson124} . The hypothalamus also activates sympathetic neurons in response to stress that induce adrenal release of epinephrine and NPY to mobilise nutrients via glycogenolysis and gluconeogenesis ^{Reference Rabasa and Dickson124} . In this context, CRF acutely produces anorexia and weight loss that may last several days ^{Reference Rabasa and Dickson124} . Indeed, either a single exposure to LPS or acute immobilisation stress reduced food intake and body weight that persisted for over a week ^{Reference Vallès, Martí, García and Armario125} . Human studies support this construct, as the most stressful event of each day increased the likelihood of eating less ^{Reference Stone and Brownell126} , which was mediated by delayed digestive processes induced by acute stress ^{Reference Wing, Blair, Epstein and McDermott127} . Acute stress in healthy adults lacking comorbidities and stressors related to obesity appear to slow digestion and reduce food intake. However, induction of inflammation during saturated HFD consumption could interact with classical stress systems to promote effects of stress on reward pathways. For example, intraperitoneal LPS injection increased plasma ACTH and corticosterone and hypothalamic/pituitary proinflammatory cytokines in CRF KO and wildtype mice ^{Reference Kariagina, Romanenko, Ren and Chesnokova128} . Reciprocally, CRF promoted cortical microglial proliferation and dose-dependently increased TNFα release mediated by phosphorylation of MAPK intracellular signalling proteins shared by TLR4 activation ^{Reference Wang, Ji and Dow129} . Therefore, HPA axis and proinflammatory molecules engage in intracellular crosstalk and can independently promote stress. Acutely, stress mobilises nutrients and dampens food intake, but these behavioural effects might be altered during up-regulated inflammatory signalling. Overall, effects of HFD intake on the interactions between inflammatory and stress systems in the NAc have not been tested to determine effects on dopamine or behaviour during obesity or diet-related stressors (Fig. 2).

The immediate effects of acute stress are contextual and influenced by environmental factors. Interestingly, NAc dopamine release was increased during cues predicting foot shock, decreased during foot shock, then stimulated after lever pressing to stop the shock ^{Reference Oleson, Gentry, Chioma and Cheer130,Reference Wenzel, Rauscher, Cheer and Oleson131} . This shows NAc dopamine neurotransmission is involved in learning behaviours to avoid negative stimuli. Stress may directly induce dopamine activity, as CRF dose-dependently increased VTA dopamine neuron firing which was abolished by antagonising CRF receptor 1 ^{Reference Wanat, Hopf, Stuber, Phillips and Bonci132} , and CRF may activate both VTA GABA and dopamine neurons ^{Reference Korotkova, Brown, Sergeeva, Ponomarenko and Haas133} . However, pharmacological activation of the HPA axis increased progressive ratio response in HFD but not chow-fed groups, which was reversed by antagonising CRF receptor 1 ^{Reference Liu134} . Importantly, acute stress does not occur in isolation, and up-regulation of inflammatory signalling by HFD consumption could alter effects of acute stress. For example, LPS application 24 h after acute, inescapable tail shock promoted hippocampal NF-κB, TNFα, IL-6 and IL-1β gene expression, corticosterone, and microglial NF-κB and IL-1β responsiveness that was fully attenuated when TLR2, TLR4 ^{Reference Weber, Frank, Sobesky, Watkins and Maier135} and glucocorticoid receptors ^{Reference Frank, Thompson, Watkins and Maier136} were blocked during stress, and this stress-induced priming of neuroinflammation was mediated by transition of microglia, but not astrocytes, from a quiescent to an active state ^{Reference Frank, Baratta, Sprunger, Watkins and Maier137} . There are also sex effects within stress-induced priming of inflammatory microglial activation. The same stress paradigm similarly primed central proinflammatory cytokines, reduced anti-inflammatory pathways, and acutely reduced sucrose intake in males and females ^{Reference Fonken, Frank and Gaudet138} . However, LPS 24 h after tail shock or glucocorticoid injection induced peripheral proinflammatory cytokines and reduced central glucocorticoid receptors in females but induced microglial IL-6 and IL-1β mRNA expression specifically in males ^{Reference Fonken, Frank and Gaudet138} . Overall, specific effects of acute stress on food intake depend on prior exposure to a stressor and complex peripheral and central inflammatory signalling, suggesting those with chronic, diet-induced elevated proinflammatory states might be primed to be more reactive and have alternate behavioural responses to chronic stress, though further study is required explore this interaction in dopaminergic centres that influence food intake, like the NAc.

Chronic stress and food intake

Chronic, repeated stressors may affect food intake and body weight differently than acute stress, particularly in an obesogenic state which may promote socio-behavioural and physiological stressors. Effects of stress on hypothalamic nuclei may override homeostatic feeding. For example, inescapable foot shock engaged the HPA axis in addition to increased NPY but decreased AGRP expression, and α-melanocyte stimulating hormone released by ARC POMC/CART neurons increased stress-induced HPA activation ^{Reference Kas, Bruijnzeel, Haanstra, Wiegant and Adan139} . Furthermore, associating a place or flavour with NPY/AGRP neuron activation reduced preference for that place or food ^{Reference Betley, Xu and Cao140} , and AGRP neuron activity was reduced due to food cues ^{Reference Betley, Xu and Cao140} and initiation of feeding ^{Reference Mandelblat-Cerf, Ramesh and Burgess141} . Therefore, negative valence associated with firing of AGRP neurons may provide motivation to perform behaviours that induce positive valence like eating ^{Reference Ferrario, Labouèbe and Liu2} which could provide a stress alleviating effect. This is important given obesity-prone rats fed HFD for 6 weeks had greater HPA axis activity with greater central CRF concentrations and plasma corticosterone than chow counterparts ^{Reference Shin, Mohankumar and Sirivelu142} . Therefore, the transition in stress response that occurs over repeated stress exposures could be exacerbated by diet-induced obesity and greater stress reactivity. Acute restraint decreased ARC AGRP expression ^{Reference Chagra, Zavala, Hall and Gosselink143} but increased hypothalamic POMC expression in conjunction with reduced food intake ^{Reference Calvez, Fromentin and Nadkarni144} . Conversely, chronic restraint increased LH AGRP and decreased ARC melanocortin 4 receptor ^{Reference Chagra, Zavala, Hall and Gosselink143} that responds to POMC. This showed unique activation of hypothalamic anorectic versus orexigenic signalling, respectively, in response to acute versus chronic stress. Diet-induced inflammation may mediate these effects, as acute lipid infusion or palmitate exposure in AGRP/NPY neurons increased TNFα, proinflammatory cytokines and NPY expression ^{Reference Dalvi, Chalmers and Luo145} . Conversely, 8-week HFD consumption reduced TNFα and AGRP expression while 20-week consumption promoted TNFα in NPY and AGRP neurons ^{Reference Dalvi, Chalmers and Luo145} . This suggests palatable foods acutely induce hypothalamic inflammation in regions that innervate the NAc followed by up-regulation of protective mechanisms which are eventually overcome with extended HFD intake, and dietary responses to chronic stress likely depend on inflammatory status related to food composition and availability associated with responsivity of NPY/AGRP and POMC/CART neurons and NAc dopamine cue learning.

Extended glucocorticoid responses may shift stress reactivity and impact hormonal regulation of energy stores. Glucocorticoids act to mobilise nutrients via gluconeogenesis ^{Reference Stumpo and Kletzien146} and adipose tissue lipolysis ^{Reference Gravholt, Dall, Christiansen, Møller and Schmitz147} during stress. However, these effects over an extended period could lead to hyperinsulinemia ^{Reference Gravholt, Dall, Christiansen, Møller and Schmitz147,Reference Nicod, Giusti, Besse and Tappy148} , hyperleptinemia ^{Reference Zakrzewska, Cusin and Stricker-Krongrad149} , and insulin ^{Reference Nicod, Giusti, Besse and Tappy148,Reference Giordano, Marzotti and Berardelli150} and leptin resistance. Indeed, 7 d of artificial glucocorticoid administration in healthy women increased food intake, plasma insulin and plasma leptin ^{Reference Uddén, Björntorp, Arner, Barkeling, Meurling and Rössner151} , while exercise decreased cortisol response to stress (i.e. cortisol reactivity) accompanied by reduced disordered eating ^{Reference Hopkins, Medina, Baird, Rosenfield, Powers and Smits152} . Maniam and Morris ^{Reference Maniam and Morris153} proposed glucocorticoid-induced stimulation of the dopamine system provides motivation to resolve stressful situations which may be ‘hijacked’ into motivation to obtain palatable foods. For example, in samples of women, presence of obesity increased cortisol reactivity ^{Reference Lorig, KieBl and Laessle154} which was associated with increased calorie consumption ^{Reference Epel, Lapidus, McEwen and Brownell155} . Downstream, glucocorticoids promote feeding by negatively regulating CRF and activating orexigenic NPY neurons ^{Reference Zakrzewska, Cusin and Stricker-Krongrad149,Reference Maniam and Morris153} that are inhibited by insulin and leptin and interconnected with the LH, NAc and VTA ^{Reference Ferrario, Labouèbe and Liu2} . Therefore, during diet-induced obesity, chronic stress promotes glucocorticoid-induced activation of feeding centres and removes the stop signal from hypothalamus to VTA to promote dopamine-motivated feeding. It is important to note that glucocorticoids are involved in normal feeding responses, as restoring corticosterone to adrenalectomised rats increased saccharin ^{Reference Bhatnagar, Bell, Liang, Soriano, Nagy and Dallman156} and sucrose ^{Reference Bell, Bhatnagar, Liang, Soriano, Nagy and Dallman157} intake to control levels. On the other hand, access to lard blunted corticosterone response during restraint stress ^{Reference La Fleur, Houshyar, Roy and Dallman158} . Therefore, diet interacts with glucocorticoids to modulate stress-induced feeding. Indeed, mice exposed to chronic social stress after 12 weeks on a HFD showed improved weight loss, reduced plasma insulin and leptin, and dampened anxio-depressive behaviours versus stressed controls or non-stressed HFD groups ^{Reference Finger, Dinan and Cryan159} , suggesting a potentially stress-alleviating effect of palatable food intake. Overall, diet-induced obesity produces neurochemical shifts and maladaptive stress responses, but an opportunity for future investigation lies in whether HPA axis and κ-opioid receptor stress responses drive food intake for obese individuals during the chronic diet-related stress of restricting calories for weight loss (Fig. 3).

Effects of a HFD on κ-opioid-receptor-mediated stress, dopamine, and food intake

The HPA axis interacts with κ-opioid receptors (KORs) to gate NAc dopamine signalling. Indeed, expression of CRF, ACTH and glucocorticoids systemically and centrally are co-regulated with dynorphin, a ligand for KORs, and its precursor prodynorphin with ligands in either system inducing expression and release reciprocally to promote stress responses ^{Reference Volkow, Wang and Baler160–Reference Wittmann, Schunk and Rosskothen163} . KORs are G-protein-coupled receptors that inhibit adenylyl cyclase and cAMP activity ^{Reference Konkoy and Childers164,Reference Prather, McGinn, Claude, Liu-Chen, Loh and Law165} and are co-expressed within NAc DAT-expressing neurons ^{Reference Svingos, Chavkin, Colago and Pickel166} . Blocking NAc KORs inhibited dopamine reuptake and increased extracellular dopamine ^{Reference Zhu and Reith167} , while KOR activation increased reuptake to limit extracellular dopamine ^{Reference Spanagel, Herz and Shippenberg168,Reference Thompson, Zapata, Justice, Vaughan, Sharpe and Shippenberg169} . KORs are present within NAc presynaptic dopamine terminals, MSNs, and GABAergic and cholinergic interneurons and reduce extracellular dopamine to promote feeding cessation ^{Reference Karkhanis, Holleran and Jones170} . However, ventricular KOR agonism promoted HFD intake during satiation, whereas KOR antagonism in a fasted state reduced HFD intake ^{Reference Ookuma, Barton, York and Bray171} , and systemic KOR agonism reduced NAc core phasic dopamine release parallel to reduced motivation to obtain sucrose ^{Reference Ebner, Roitman, Potter, Rachlin and Chartoff172} . This suggests KORs control extracellular dopamine to alter rewarding effects of food. In the NAc, KOR activation during nicotine exposure ^{Reference Tejeda, Natividad, Orfila, Torres and O’Dell173} and amphetamine withdrawal ^{Reference Del Pilar Escobar, Cornejo, Andrés and Fuentealba174} decreased basal and evoked dopamine, while KOR antagonism attenuated alcohol self-administration during withdrawal ^{Reference Nealey, Smith, Davis, Smith and Walker175} . Furthermore, acute stress promoted TNFα in the amygdala, and amygdalar TNFα potentiated anxiety during withdrawal that was blocked by CRF antagonism ^{Reference Knapp, Whitman and Wills176} . Therefore, HPA axis, KOR and inflammatory systems interact to induce dysphoria and relapse during withdrawal, and up-regulation of these symptoms during diet-induced obesity could promote food cravings when energy is restricted for weight loss.

Neuroinflammation interacts with stress and reward systems. Indeed, exposing women to stress induced plasma IL-6 concentrations that correlated with reduced NAc activation during an fMRI reward learning task ^{Reference Treadway, Admon and Arulpragasam177} . Further, chronic intake of saturated HFD versus low fat ^{Reference Décarie-Spain, Sharma and Hryhorczuk118,Reference Dutheil, Ota, Wohleb, Rasmussen and Duman121,Reference Aslani, Vieira, Marques, Costa, Sousa and Palha178} or monounsaturated fat ^{Reference Décarie-Spain, Sharma and Hryhorczuk118} increased anxio-depressive behaviours, plasma corticosterone, and hypothalamic and NAc inflammatory cytokine and TLR4 expression. Conversely, prodynorphin overexpression promoted anti-inflammatory M2 versus M1 phenotype of hippocampal microglia and inhibited LPS-induced TLR4 activation of NF-κB to promote IL-4 and IL-10 but inhibit IL-1β and IL-6 release ^{Reference Liu, Xu, Dai, Tan, Mao and Chen179} . Reciprocally, intraplantar injection of IL-1β increased KOR mRNA expression in basal root ganglia neurons ^{Reference Puehler, Rittner and Mousa180} . However, dynorphin, enkephalin and KOR expression in the NAc were not affected by WD intake or 18-d replacement with chow ^{Reference Alsiö, Olszewski and Norbäck181} or 18 d of ad libitum or intermittent access to a HFD ^{Reference Bake, Baron, Duncan, Morgan and Mercer182} . Therefore, while HFD intake itself does not appear to alter dynorphin/KOR expression, it appears that KORs dampen proinflammatory responses induced by saturated fat and could engage in intracellular crosstalk with inflammatory molecules similarly to the HPA axis. Human neuroimaging studies linked a dopaminergic response to the consumption of palatable food ^{Reference Small, Zatorre, Dagher, Evans and Jones-Gotman183} , palatable food cues ^{Reference Eisenstein, Black and Samara29} or smelling food odours ^{Reference Frasnelli, Hummel, Bojanowski, Warr, Gerber and Hummel184} with increased reports of pleasure ^{Reference Small, Zatorre, Dagher, Evans and Jones-Gotman183,Reference Frasnelli, Hummel, Bojanowski, Warr, Gerber and Hummel184} and wanting ^{Reference Eisenstein, Black and Samara29} of food. Therefore, changing a lifetime of food preferences by restricting energy intake to lose weight may be particularly stressful and activate stress pathways that dampen dopamine signalling.

In support of KOR system involvement during palatable food restriction, pair feeding of WD to chow intake levels in controls reduced NAc KOR expression in absence of obesity, but 18 d of WD replacement with chow did not alter KOR expression ^{Reference Alsiö, Olszewski and Norbäck181} . Therefore, it is possible that differences in KOR expression become neutralised over 18 d of diet replacement or by diet-induced weight gain, or that specifically KOR function could be up-regulated. HFD intake alters food preferences, as replacement of a WD ^{Reference Pickering, Alsiö, Hulting and Schiöth45} or HFD ^{Reference Sharma, Fernandes and Fulton8} with chow persistently reduced food intake and acutely increased plasma corticosterone and anxiety-like behaviour ^{Reference Sharma, Fernandes and Fulton8} . Further, whereas HFD intake promoted sucrose anhedonia ^{Reference Sharma, Fernandes and Fulton8,Reference Carlin, Hill-Smith, Lucki and Reyes48,Reference Carlin, McKee and Hill-Smith185} , palatable food removal increased preference for sucrose ^{Reference Carlin, Hill-Smith, Lucki and Reyes48,Reference Carlin, McKee and Hill-Smith185} and motivation to obtain sucrose ^{Reference Sharma, Fernandes and Fulton8,Reference Pickering, Alsiö, Hulting and Schiöth45} and fat ^{Reference Sharma, Fernandes and Fulton8} that occurred as soon 3 d ^{Reference Sharma, Fernandes and Fulton8} after diet removal but persisted for several weeks ^{Reference Pickering, Alsiö, Hulting and Schiöth45,Reference Carlin, Hill-Smith, Lucki and Reyes48,Reference Carlin, McKee and Hill-Smith185} . This indicates prolonged enhancement of salience for palatable food when a preferred diet is not available, and it is possible that KORs, which contribute to the rewarding properties of food, promote this effect. Further, HFD intake started at parturition reduced VTA TH and NAc D1R and D2R expression in male and female mice, but 4 weeks of HFD replacement exacerbated this reduction in the VTA of females and NAc of males but oppositely restored these effects in the VTA of males and NAc of females ^{Reference Carlin, Hill-Smith, Lucki and Reyes48} . However, when HFD is introduced in adolescence, HFD replacement does not recover D2R expression in females with strong promotion of palatable food intake, though effects were attenuated during adult-onset HFD ^{Reference Carlin, McKee and Hill-Smith185} . This suggests prolonged effects of a HFD on dopamine synthesis and signalling in the VTA alter the capacity for neurotransmission in the NAc. Changes in NAc neurotransmission occur via alterations to D1Rs and D2Rs resulting in increased salience of palatable foods when they are no longer available. These effects were significantly mediated by sex and age of palatable food introduction with high vulnerability during adolescence. Overall, individuals with clinical obesity or that persistently consume a HFD may be poorly adapted to control palatable food cravings during energy restriction owing to diet-induced or psychological stress that engages the HPA or KOR systems.