- Email: [email protected]
Food motivated behavior of melanocortin-4 receptor knockout mice under a progressive ratio schedule
peptides 27 (2006) 2829–2835
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/peptides
Food motivated behavior of melanocortin-4 receptor knockout mice under a progressive ratio schedule C. Vaughan a, M. Moore b, C. Haskell-Luevano b, N.E. Rowland a,* a b
Department of Psychology, University of Florida, Gainesville, FL 32611, United States Department of Medicinal Chemistry, University of Florida, Gainesville, FL 32611 United States
article info
abstract
Article history:
Melanocortin-4 receptor knockout (MC4RKO) mice are hyperphagic and develop obesity
Received 16 May 2006
under free feeding conditions. We reported previously that MC4RKO mice did not maintain
Received in revised form
hyperphagia and as a result lost weight when required to press a lever to obtain food on a
7 July 2006
fixed ratio procurement schedule. To assess the generality of this result, we tested MC4RKO
Accepted 12 July 2006
mice and their heterozygous and wild type littermates using progressive ratio (PR) schedules
Published on line 22 August 2006
that are believed to be sensitive indicators of motivation. Mice lived in operant chambers and obtained all of their food (20 mg pellets) via lever press responding. Food was available
Keywords:
according to a PR schedule so that within a meal, food became progressively more costly,
Melanocortin-4 receptors
and we expected this would provide a stringent test of mechanisms controlling meal size.
Knockout
The schedule reset after either 3 or 20 min of no responding, so defining meals, and the
Mice
highest ratio completed before the reset was defined as the breakpoint. The average daily
Operant conditioning
number of meals was lower and mean size of meals was higher at the 20 compared with the
Progressive ratio
3 min reset condition. Mean daily food intake did not differ between the two reset criteria
Closed economy
but did differ as a function of genotype, with MC4RKO mice eating about 25% more than
Food intake
heterozygous or wild type mice. Hyperphagia in the MC4RKO mice was characterized primarily by larger meals (higher breakpoints) and they emitted about twice as many responses as wild type mice. Thus, using a PR schedule, MC4RKO mice exhibit hyperphagia, and show a high level of motivation to support large meal sizes. # 2006 Elsevier Inc. All rights reserved.
1.
Introduction
The obesity epidemic has been a topic of interest recently due to the rise in obesity-related illnesses. The sharp increase in the incidence of humans identified as overweight in the past 10–20 years is directly attributable to environmental and lifestyle changes. Physiologic and genomic research using animal models have additionally identified genes and polymorphisms that predispose carriers to obesity. The melanocortin system has received recent attention both in the human domain as well as in animal models including genetically engineered mice (see [11] for review). The two central
melanocortin receptors (MCRs) that are most pertinent to feeding are the MC3R and the MC4R (see [4,25] for review). Functionally, the MC3R is related to fat metabolism while the MC4R is related to food intake and energy expenditure [2,6,9,12,15]. There is emerging evidence that feeding in the MC4R knockout (KO) mouse model is subject to environmental modulation [21,23]. Specifically, while hyperphagia is essential to maintain their obese state [23], MC4RKO mice did not become obese when a running wheel was present but they developed hyperphagia and obesity when housed individually without a wheel [16]. We have also reported a protocol in
* Corresponding author. Tel.: +1 352 392 0601; fax: +1 352 392 7985. E-mail address: [email protected] (N.E. Rowland). 0196-9781/$ – see front matter # 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.peptides.2006.07.008
2830
peptides 27 (2006) 2829–2835
which MC4RKO mice were not hyperphagic, and repeatedly lost weight, when required to work for food in operant chambers [24]. However, each time they were returned to free feeding, their weight rebounded, presumably due to hyperphagia. The operant protocol that we used was one of imposed procurement costs, when to initiate a meal a fixed ratio (FR) of lever presses had to be emitted. Both wild type and MC4RKO mice showed changes in meal size and frequency as a result of the imposed FR, but the KO mice ate the same amount as the wild types. One interpretation of this finding is that MC4RKO mice are less motivated to work for food, so in the present study we apply a stringent test of this hypothesis using progressive ratio (PR) schedules that are thought to be sensitive indicators of motivation for either drugs or food [13,14,18,22]. PR schedules require a subject to emit an increasing number of responses to gain access to successive reinforcers [13]. Typically, these are used in short sessions that are terminated when the subject will not emit the work needed to obtain another reinforcer, and this is termed the breakpoint: greater motivation produces higher breakpoints [10,13]. We have adapted the PR protocol to measure meal patterns by using a closed economy in which mice live and obtain all of their food. To do this, we have employed a reset condition; if there is an elapsed criterion time without responding the response requirement reverts to the initial value (one press) and the animal gains the opportunity to initiate a new meal. Hence, the animal has complete control over the initiation and termination of meals. Due to the fact that the first pellets in any meal are the cheapest, a strategy of many small meals will minimize the average cost of food (presses per pellet). If, as we suggested above, MC4RKO mice are less motivated than wild types, then we would expect they will eat smaller meals, and depending on their compensation in intermeal interval, in a day they may eat either less, the same, or more than wild type mice.
2.
Materials and methods
2.1.
Subjects and housing environment
Untimed pregnant heterozygous mothers were from a colony originating from Millennium Pharmaceuticals and maintained by one of us (CHL) at the University of Florida. Mothers gave birth 3–19 days after being moved to a vivarium in the Psychology department. As part of another study, these offspring were reared in small (4 pups) or larger (8 pups) litters and were fed Chow or high fat diet until weaned and were maintained on these diets until 77 days of age when they were returned to Chow for at least 12 weeks before the present work. These manipulations had only small effects on body weight, and the offspring were evenly distributed among the genotypes so there would be no systematic effect on the results of the present experiment. Five MC4RKO, five heterozygotes (HET) and five wild type (WT) adult male mice were used in the current experiment. Only males were used to avoid the variation in daily intake and meal patterns that might result from estrous cyclicity. Mice were genotyped using PCR to amplify DNA from a tail snip
taken between 8 and 10 weeks of age. The vivarium was illuminated from 08:00 to 20:00 h and was maintained at an ambient temperature of 23 2 8C. Mice were housed individually in standard shoebox cages with water and food (Purina 5001 Chow) available ad libitum. Mice were then familiarized with the experimental food by presenting a jar of 20-mg nutritionally complete food pellets (Noyes Precision pellets, Research Diets Inc., New Brunswick NJ; 25% protein, 64% carbohydrate, 11% fat) in home cages for 1 day. Next, mice were placed into operant chambers (13 cm 13 cm 12 cm) for a magazine training session for 1 day. During the session, a pellet was automatically delivered every 8 min for 24 h. Water was available from a spout in the middle of the wall opposite the levers and food trough. The chambers were contained inside ventilated, sound attenuating cubicles, each with a 15-W light providing the same 12:12 cycle as the vivarium. Cue lights, real time records of pellet delivery and lever presses were controlled by Med-PC software (MED Associates Inc., St. Albans, VT). During all operant sessions, mice lived in the chambers with the exception of 30 min in the middle of each day when they were removed to a holding cage without food while the chambers were cleaned and serviced. Over the course of the next 3–5 days, mice received food contingent upon completion of a fixed ratio schedule; the ratio was increased (FR1 ! FR2 ! FR3) when mice showed sufficient lever pressing to maintain normal food intake. The mice were then studied at FR3 for another 7 days to establish baseline intakes. This phase of the experiment was conducted to establish whether mice would work for food; all mice performed well and were then moved to the PR schedule. Thereafter, because only four chambers were available, mice were run in squads each containing at least one mouse of each genotype.
2.2.
Progressive ratio sessions
On the PR1 schedule, the required responses for each successive pellet within a defined meal increased arithmetically (1, 2, 3, . . ., n). Each 23.5 h session began with the left cue and house lights on. Pressing could begin immediately and there was no additional cue for pellet delivery (i.e. no time out period). Whenever there was a 3-min stretch of no responding, the program reset. Resetting returned the lever press requirement to 1 and the animal was able to initiate a new period of pressing whenever desired. This reset criterion was studied for 12–13 days after which time mice were given free food in their home cage for 9–14 days. They were then returned to the chambers for another 12–13 days, but with a reset criterion of 20-min. In this situation, each session began with the cue and house lights off for 1 min. After the 1 min, the cue light was illuminated and mice could perform their PR run on this lever. During pellet delivery the house light came on briefly and was otherwise off. Whenever 20 min elapsed with no responding, the program reset. At the 20-min reset marker, a new run was signaled to the animal by a 1 min time out without cue light illumination. For both reset conditions, the number of responses, total pellets obtained, number of resets (defining meals), and average break point (defining meal size) was computed for
peptides 27 (2006) 2829–2835
2831
Table 1 – Mean daily food intake (g) during phases of experiment Phase FR3 PR 3-min reset PR 20-min reset
Wild type (n = 5)
Heterozygous (n = 5)
Knockout (n = 5)
4.2 0.1 4.0 0.1 4.2 0.1
4.6 0.1y 4.5 0.1 * 4.7 0.1 *
5.6 0.1 ** 5.2 0.1 ** 5.7 0.1 **
Values in table are the mean S.E. **Intake greater than HET and WT mice within phase ( p < 0.001). *Intake greater than WT within phase ( p < 0.001). yIntake greater than WT within phase ( p < 0.05).
each subject and session. Body weights were recorded at the beginning, middle, and end of each block of PR testing.
2.3.
Data analysis
The individual mean for total intake per day in each reset condition was calculated using 12 consecutive days of stable performance with correction for food spillage. Food spillage was counted as the number of pellets found beneath the grill floor, and this amount was subtracted from the daily intake. ANOVAs and Student t tests were used to analyze data and genotype was used as the grouping variable. The significance criterion was set at p < 0.05.
3.
Results
The mean intakes of food for each genotype in each phase of the experiment are shown in Table 1. In the FR3 condition, MC4RKO mice ate significantly more than the other two genotypes ( p < 0.001). This pattern was also present during the two PR phases, but in these phases the heterozygotes also showed a modest hyperphagia. The KO mice initially were obese, weighing 50% more than wild types and because of their sustained hyperphagia, their elevated weight was maintained throughout the experiment (Fig. 1 and Table 1). The average weight of the heterozygotes was slightly higher than the wild types. The mean responses per feeding bout from the PR phases are shown in Fig. 2. Also, shown above each bar is the mean number of pellets mice received before a breakpoint in responding. Breakpoints were lower during the 3 min than
Fig. 1 – Weights of mice during experimental phases. Group means are represented. KO > HET > WT, p < 0.001.
Fig. 2 – Mean responses per feeding bout for WT, HET and MC4RKO mice. Bars represent the mean W S.E. responses emitted per feeding bout including presses that did not directly result in pellet delivery. Numbers above each bar represent the mean number of pellets earned before a breakpoint: (A) shows responses in the 3-min reset criterion and (B) shows responses in the 20-min reset criterion. In (A) *MC4RKO mice had the highest responses per bout in comparison to WT ( p < 0.001) and HET ( p < 0.05). #HET and WT mice significantly differed ( p < 0.001) from each other. In (B) all genotypes significantly differed from each other (*KO > #HET > WT; p < 0.001).
the 20 min reset, and showed a significant effect of genotype with the highest breakpoints in the MC4RKO mice. On average, breakpoints were about three-fold higher in the 20 min compared with the 3 min condition, and were about two-fold higher in KO compared with wild type mice. The mean number of feeding bouts initiated during each PR phase is shown in Fig. 3. Mice took significantly fewer feeding bouts under the 20 min compared with the 3 min criterion, and KOs took slightly fewer feeding bouts than wild type. The behavior under the 3 min reset was essentially snacking, with about 40 feeding episodes per 23.5 h. The sum of the daily breakpoints across resets yields the total number of responses emitted per day, and these data are
2832
peptides 27 (2006) 2829–2835
Fig. 3 – Mean W S.E. number of snacks or small feeding bouts initiated by the mice under both PR1 conditions as denoted on the y-axis. Graph A represents data from the 3min reset condition. *Number of snacks were greater than HET and KO groups ( p < 0.001). Graph B represents data from the 20-min reset condition. *Number of snacks were greater than HET ( p < 0.05) and KO ( p < 0.001). #Number of snacks were greater than KO ( p < 0.05).
shown in Fig. 4. Total responses were greater at the 20 min reset than at the 3 min reset, and the MC4RKO mice emitted about two-fold more responses that the wild types, with the heterozygotes intermediate. The average price per pellet in the different reset conditions is shown in Table 2. KO and HET mice paid a higher average price per pellet WT and this effect was more pronounced in the 20-min reset phase. Fig. 5 shows 1 day of representative data for three mice, one from each genotype. Since mice were run in different squads, the start time (denoted as 0 on the x-axis) was approximately the same. These data show that more food was consumed at night compared with the daytime in all genotypes, although this was not formally analyzed.
4.
Discussion
To our knowledge this is the first report using a PR schedule in a closed economy to study meal patterns. PR schedules have been used predominantly in short sessions where breakpoints end the session, but in our case they defined elective meal
Fig. 4 – Twelve days average of total responses made per daily session. Graph A shows data from the 3-min reset criterion condition. *KO mice had a higher amount of responses daily than HET ( p < 0.01) and WT ( p < 0.001). # HET mice pressed more than WT ( p < 0.001). Graph B shows data from the 20-min reset criterion condition. *KO mice had a higher number of responses daily than HET and WT ( p < 0.001). #HET mice pressed more than WT ( p < 0.001) during this condition as well.
termination and so our interpretation of the breakpoint was most relevant to the assessment of the subjects’ ability to sense fullness and end a discrete bout of feeding. Under these conditions, the PR schedule may emulate depletion of a patch of finite resources, and the breakpoint represents a decision to move to a new patch. Additionally, PR schedules are thought to be particularly sensitive measures of motivation because they reveal the maximum price an individual is willing to pay for a fixed reinforcer [8,20]. The main purpose of the present study was to test a hypothesis that MC4RKO mice may be less motivated to work for food based on observations from a previous study [24]. In that study mice had a two lever contingency for pellet delivery; they were required to emit a certain number of responses (the procurement cost) on the left lever in order to for the right (pellet delivery) lever to be activated. At that point, a small fixed number of presses (5 or 10, the consumatory cost) on the
peptides 27 (2006) 2829–2835
Table 2 – Price (presses/number of pellets) averaged across 12 days of PR1 phases Phase PR 3-min reset PR 20-min resety
Wild type (n = 5)
Heterozygous (n = 5)
Knockout (n = 5)
3.8 0.1 5.6 0.1
5.2 0.1 ** 6.1 0.2 **
5.0 0.1 *** 7.7 0.2 ***
Values in table are the mean S.E. ***Price greater than HET and WT mice within phase ( p < 0.001). **Price greater than WT within phase ( p < 0.001). yOverall price greater than 3 min reset phase.
right lever delivered single pellets. However, after 10 min of no pressing on the right lever, the program de-activated this lever and in order to again obtain food again the animal had to run off a new procurement cost on the left lever. In both WT and MC4RKO mice, the number of meals per day declined but their size increased as the procurement cost rose, showing that the animals avoided work by changing their meal strategy. At the lowest cost on both levers, mice emitted 1000 presses per day and intakes were comparable for WT (3.6 g) and MC4RKO (4.0 g), and as a result the MC4RKO mice lost weight.
Fig. 5 – Cumulative responses of individual mice across a daily session; the actual start time for each session is indicated in the legend. Graph A shows cumulative responses for the 3-min reset condition. Graph B shows cumulative responses for the 20-min reset condition. Stepwise response increments are steeper during the 20-min reset condition for these three mice.
2833
In the present experiment, the results were strikingly different. First, in the training phase using an FR3 (consumatory cost only), MC4RKO mice were hyperphagic relative to WT. Critical differences from our previous study include the fact there was only one lever, and no ‘penalty’ for time without activity. These and other differences require further analysis. Our present results during the PR phases are inconsistent with our hypothesis and actually suggest the reverse. We expected that mice would adopt a strategy of small meals and frequent switching because the passage of time (reset criterion) was the only ‘‘cost’’ associated with initiating a new meal. This was especially evident at the lower (3 min) reset time. In part, we expected this strategy because previous studies using shorter sessions found differences between performance of ad libitum fed rodents and performance of rodents kept at 80% of their free feeding weight [14,17]. Our mice were not food deprived in this paradigm and thus determined their own pattern of feeding bouts. The low number of responses per feeding bout and frequent initiation of small bouts we observed during the 3min reset could in part also have been due to the lack of cue signaling. This may have confused the mice, which is why we added a cue and time out structure for the 20 min reset. Thus, the higher number of responses per feeding bout seen for all mice at this reset could have been due to the time, the cueing, or both. Future studies will be needed to examine in more detail the parameters of this protocol. Contrary to our hypothesis, KO mice did not engage in small, cheap meals but instead worked on average more per pellet than the other genotypes (Table 2). As proposed earlier as a possible outcome, we did see compensation by MC4RKO mice; their intermeal intervals were longer in the 3-min condition versus the 20 min condition (Fig. 5). Therefore, KO mice achieved hyperphagia under our PR1 schedules by having significantly higher meal sizes, as defined by number of responses per feeding bout, and lower meal frequencies, as defined by number of small feeding bouts, than the other genotypes. Breakpoints were higher for MC4RKO mice in the 20-min reset condition in comparison to the 3-min reset condition. At an average of 111 47 presses before taking a break, KO mice received about 14 pellets (0.28 g) per defined meal. During this period KO mice were initiating feeding bouts around 22 0.6 times per day. At the highest number of responses emitted per feeding bout of 158 presses, KO mice show an estimated breakpoint of about 17 pellets (0.34 g) in that discrete period of pressing. There was a reduction in the number of small feeding bouts seen for all mice to an average of 22–25 bouts per day during the 20-min reset condition. This shows that time available for feeding plays an important role in feeding strategy and in conjunction with the data on responses emitted (Fig. 4) suggest that the MC4R plays no role in adjusting effort to time constraints under these conditions. Due to the hyperphagic phenotype, it could be said that the KO mice were internally driven to maximize food intake, especially in the 20 min reset condition. Further evidence shows that MC4RKO mice paid a higher price per pellet than WT mice (Table 2) and thus can be considered more motivated than WT mice.
2834
peptides 27 (2006) 2829–2835
We examined the data for genotype-specific 24-h feeding patterns by analyzing records from three representative mice (Fig. 5). The three mice showed different profiles under the 3min versus the 20-min criterion. The representative WT nearly ceases feeding in the early morning hours of both reset conditions, consistent with previous work in mice [1]. The time constraint imposed on mice in the 20 min reset condition resulted in an increased rate of pressing for all genotypes. The KO and HET mice continue to lever press and eat well into the morning hours (Fig. 5). In an early study, Anliker and Mayer [1] found that ob/ob mice on an FR 25 schedule had very high response rate. ob/ob mice were hyperphagic on this FR schedule and showed no cyclical day/night pattern of eating. A similar consistent eating pattern characteristic of hyperphagia was also seen in the MC4RKO mouse in the present study (Fig. 5). Collier and colleagues [7] have shown that under two lever FR schedules in which work can be minimized by taking fewer larger meals, rats and many other species adopt exactly this strategy. It is thus not surprising that the converse of this pattern can be found under PR conditions in which work is avoidable by initiating small feeding bouts. We observed no decrease in weight when mice were in the chambers pressing for food. This suggests the effort of lever pressing did not contribute a significant metabolic demand. In past reports, hyperphagia in MC4RKO mice tends to be highly influenced by the housing environment [5,16,24]. It has been established that KO mice are missing an important receptor for the control of food intake but we still do not fully know what effect this has on postingestive feedback. The HET and WT mice may have more accurate negative feedback to end meals in comparison to KO mice and this may be most evident when mice have to work for food. HET and WT mice ate less than KOs but still maintained their body weights overall; this could be due to a number of external and/or internal factors. In summary, the deletion of the MC4R leads to increased food intake in mice [5,15,16,21,23]. Our previous experiment [24] did not find MC4RKO mice to be hyperphagic in the chamber, however when tested in a 24-h PR schedule MC4RKOs did exhibit significant and showed evidence of planning feeding bouts to maximize the amount of effort that contributed to feeding. The present experiment also introduces a novel protocol to study meal patterns: a closed economy PR protocol. We expected that this protocol would be a particularly stringent test of large meal phenotypes, and requires further validation using other protocols including for example chronic treatment with pharmacological agents that alter food intake. Through this study and others it is evident the MC4R gene positively influences descending information regarding feeding related behaviors. Though the MC4R is located in brain areas ascribed to be involved in motivation [3,19], the involvement of the MC4R seems to have the most significant role in the quantitative aspects of meal taking, like how much is eaten and how many calories to expend to gain access to food. This PR experiment has successful applied parameters of behavioral economics to MC4RKO feeding. Such an approach is useful tool in discovering how subjects adjust meal taking to different simulated environments.
Acknowledgements The MC4RKO mice were generously provided by Millennium Pharmaceuticals and this work was supported in part by NIHDK57080 (CHL).
references
[1] Anliker J, Mayer J. An operant conditioning technique for studying feeding-fasting patterns in normal and obese mice. J Appl Physiol 1956;8:667–70. [2] Balthasar N, Dalgaard LT, Lee CE, Yu J, Funahashi H, Williams T, et al. Divergence of melanocortin pathways in the control of food intake and energy expenditure. Cell 2005;123:493–505. [3] Berthoud HR. Multiple neural systems controlling food intake and body weight. Neurosci Biobehav Rev 2005;26:393–428. [4] Butler AA, Cone RD. Knockout studies defining different roles for melanocortin receptors in energy homeostasis. Ann NY Acad Sci 2003;994:240–5. [5] Butler AA, Marks DL, Fan W, Kuhn CM, Bartolome M, Cone RD. Melanocortin-4 receptor is required for acute homeostatic responses to increased dietary fat. Nat Neurosci 2001;4:605–11. [6] Chen AS, Metzger JM, Trumbauer ME, Guan X, Yu H, Frazier EG, et al. Role of the melanocortin-4 receptor in metabolic rate and food intake in mice. Transgenic Res 2000;9:145–54. [7] Collier GH, Johnson DF, Hill WL, Kaufman LW. The economics of the law of effect. J Exp Anal Behav 1986;46:113–36. [8] El-Ghundi M, O’Dowd BF, Erclik M, George SR. Attenuation of sucrose reinforcement in dopamine D1 receptor deficient mice. Eur J Neurosci 2003;17:851–62. [9] Farooqi IS, Keogh JM, Yeo GS, Lank EJ, Cheetham T, O’Rahilly S. Clinical spectrum of obesity and mutations in the melanocortin-4 receptor gene. N Engl J Med 2003;348:1096–103. [10] Findley JD. Preference and switching under concurrent scheduling. J Exp Anal Behav 1958;1:123–44. [11] Hadley ME, Haskell-Luevano C. The proopiomelanocortin system. Ann NY Acad Sci 1999;885:1–21. [12] Harrold JA, Widdowson PS, Williams G. Altered energy balance causes selective changes in melanocortin-4 (MC4R), but not melanocortin-3 (MC3-R), receptors in specific hypothalamic regions: further evidence that activation of MC4-R is a physiological inhibitor of feeding. Diabetes 1999;48:267–71. [13] Hodos W. Progressive ratio as a measure of reward strength. Science 1961;134:943–4. [14] Hodos W, Kalman G. Effects of increment size and reinforcer volume on progressive ration performance. J Exp Anal Behav 1963;6:387–92. [15] Huszar D, Lynch CA, Fairchild-Huntress V, Dunmore JH, Fang Q, Berkemeier LR, et al. Targeted disruption of the melanocortin-4 receptor results in obesity in mice. Cell 1997;88:131–41. [16] Irani BG, Xiang Z, Moore MC, Mandel RJ, Haskell-Luevano C. Voluntary exercise delays monogenetic obesity and overcomes reproductive dysfunction of the melanocortin-4 receptor knockout mouse. Biochem Biophys Res Commun 2005;326:638–44. [17] Jewett DC, Cleary J, Levine AS, Schaal DW, Thompson T. Effects of neuropeptide Y, insulin, 2-deoxyglucose, and food deprivation on food-motivated behavior. Psychopharmacology 1995;120:267–71.
peptides 27 (2006) 2829–2835
[18] Low MJ, Hayward MD, Appleyard SM, Rubinstein M. Statedependent modulation of feeding behavior by proopiomelanocortin-derived beta-endorphin. Ann NY Acad Sci 2003;994:192–201. [19] Mountjoy KG, Mortrud MT, Low MJ, Simerly RB, Cone RD. Localization of the melanocortin-4 receptor (MC4-R) in neuroendocrine and autonomic control circuits in the brain. Mol Endocrinol 1994;8:1298–308. [20] Richardson NR, Roberts DCS. Progressive ratio schedules in drug self-administration studies in rats: a method to evaluate reinforcing efficacy. J Neurosci Meth 1996;66:1–11. [21] Samama P, Rumennik L, Grippo JF. The melanocortin receptor MCR4 controls fat consumption. Regul Pept 2003;113:85–8.
2835
[22] Stafford D, LeSage MG, Glowa JR. Progressive-ratio schedules of drug delivery in the analysis of drug selfadministration: a review. Psychopharmacology 1998;139:169–82. [23] Ste Marie L, Miura GI, Marsh DJ, Yagaloff K, Palmiter RD. A metabolic defect promotes obesity in mice lacking melanocortin-4 receptors. Proc Natl Acad Sci USA 2000;97:12339–44. [24] Vaughan CH, Moore MC, Haskell-Luevano C, Rowland NE. Meal patterns and foraging in melanocortin receptor knockout mice. Physiol Behav 2005;84:129–33. [25] Vergoni AV, Schioth HB, Bertolini A. Melanocortins and feeding behavior. Biomed Pharmacother 2000;54: 129–34.
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/peptides
Food motivated behavior of melanocortin-4 receptor knockout mice under a progressive ratio schedule C. Vaughan a, M. Moore b, C. Haskell-Luevano b, N.E. Rowland a,* a b
Department of Psychology, University of Florida, Gainesville, FL 32611, United States Department of Medicinal Chemistry, University of Florida, Gainesville, FL 32611 United States
article info
abstract
Article history:
Melanocortin-4 receptor knockout (MC4RKO) mice are hyperphagic and develop obesity
Received 16 May 2006
under free feeding conditions. We reported previously that MC4RKO mice did not maintain
Received in revised form
hyperphagia and as a result lost weight when required to press a lever to obtain food on a
7 July 2006
fixed ratio procurement schedule. To assess the generality of this result, we tested MC4RKO
Accepted 12 July 2006
mice and their heterozygous and wild type littermates using progressive ratio (PR) schedules
Published on line 22 August 2006
that are believed to be sensitive indicators of motivation. Mice lived in operant chambers and obtained all of their food (20 mg pellets) via lever press responding. Food was available
Keywords:
according to a PR schedule so that within a meal, food became progressively more costly,
Melanocortin-4 receptors
and we expected this would provide a stringent test of mechanisms controlling meal size.
Knockout
The schedule reset after either 3 or 20 min of no responding, so defining meals, and the
Mice
highest ratio completed before the reset was defined as the breakpoint. The average daily
Operant conditioning
number of meals was lower and mean size of meals was higher at the 20 compared with the
Progressive ratio
3 min reset condition. Mean daily food intake did not differ between the two reset criteria
Closed economy
but did differ as a function of genotype, with MC4RKO mice eating about 25% more than
Food intake
heterozygous or wild type mice. Hyperphagia in the MC4RKO mice was characterized primarily by larger meals (higher breakpoints) and they emitted about twice as many responses as wild type mice. Thus, using a PR schedule, MC4RKO mice exhibit hyperphagia, and show a high level of motivation to support large meal sizes. # 2006 Elsevier Inc. All rights reserved.
1.
Introduction
The obesity epidemic has been a topic of interest recently due to the rise in obesity-related illnesses. The sharp increase in the incidence of humans identified as overweight in the past 10–20 years is directly attributable to environmental and lifestyle changes. Physiologic and genomic research using animal models have additionally identified genes and polymorphisms that predispose carriers to obesity. The melanocortin system has received recent attention both in the human domain as well as in animal models including genetically engineered mice (see [11] for review). The two central
melanocortin receptors (MCRs) that are most pertinent to feeding are the MC3R and the MC4R (see [4,25] for review). Functionally, the MC3R is related to fat metabolism while the MC4R is related to food intake and energy expenditure [2,6,9,12,15]. There is emerging evidence that feeding in the MC4R knockout (KO) mouse model is subject to environmental modulation [21,23]. Specifically, while hyperphagia is essential to maintain their obese state [23], MC4RKO mice did not become obese when a running wheel was present but they developed hyperphagia and obesity when housed individually without a wheel [16]. We have also reported a protocol in
* Corresponding author. Tel.: +1 352 392 0601; fax: +1 352 392 7985. E-mail address: [email protected] (N.E. Rowland). 0196-9781/$ – see front matter # 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.peptides.2006.07.008
2830
peptides 27 (2006) 2829–2835
which MC4RKO mice were not hyperphagic, and repeatedly lost weight, when required to work for food in operant chambers [24]. However, each time they were returned to free feeding, their weight rebounded, presumably due to hyperphagia. The operant protocol that we used was one of imposed procurement costs, when to initiate a meal a fixed ratio (FR) of lever presses had to be emitted. Both wild type and MC4RKO mice showed changes in meal size and frequency as a result of the imposed FR, but the KO mice ate the same amount as the wild types. One interpretation of this finding is that MC4RKO mice are less motivated to work for food, so in the present study we apply a stringent test of this hypothesis using progressive ratio (PR) schedules that are thought to be sensitive indicators of motivation for either drugs or food [13,14,18,22]. PR schedules require a subject to emit an increasing number of responses to gain access to successive reinforcers [13]. Typically, these are used in short sessions that are terminated when the subject will not emit the work needed to obtain another reinforcer, and this is termed the breakpoint: greater motivation produces higher breakpoints [10,13]. We have adapted the PR protocol to measure meal patterns by using a closed economy in which mice live and obtain all of their food. To do this, we have employed a reset condition; if there is an elapsed criterion time without responding the response requirement reverts to the initial value (one press) and the animal gains the opportunity to initiate a new meal. Hence, the animal has complete control over the initiation and termination of meals. Due to the fact that the first pellets in any meal are the cheapest, a strategy of many small meals will minimize the average cost of food (presses per pellet). If, as we suggested above, MC4RKO mice are less motivated than wild types, then we would expect they will eat smaller meals, and depending on their compensation in intermeal interval, in a day they may eat either less, the same, or more than wild type mice.
2.
Materials and methods
2.1.
Subjects and housing environment
Untimed pregnant heterozygous mothers were from a colony originating from Millennium Pharmaceuticals and maintained by one of us (CHL) at the University of Florida. Mothers gave birth 3–19 days after being moved to a vivarium in the Psychology department. As part of another study, these offspring were reared in small (4 pups) or larger (8 pups) litters and were fed Chow or high fat diet until weaned and were maintained on these diets until 77 days of age when they were returned to Chow for at least 12 weeks before the present work. These manipulations had only small effects on body weight, and the offspring were evenly distributed among the genotypes so there would be no systematic effect on the results of the present experiment. Five MC4RKO, five heterozygotes (HET) and five wild type (WT) adult male mice were used in the current experiment. Only males were used to avoid the variation in daily intake and meal patterns that might result from estrous cyclicity. Mice were genotyped using PCR to amplify DNA from a tail snip
taken between 8 and 10 weeks of age. The vivarium was illuminated from 08:00 to 20:00 h and was maintained at an ambient temperature of 23 2 8C. Mice were housed individually in standard shoebox cages with water and food (Purina 5001 Chow) available ad libitum. Mice were then familiarized with the experimental food by presenting a jar of 20-mg nutritionally complete food pellets (Noyes Precision pellets, Research Diets Inc., New Brunswick NJ; 25% protein, 64% carbohydrate, 11% fat) in home cages for 1 day. Next, mice were placed into operant chambers (13 cm 13 cm 12 cm) for a magazine training session for 1 day. During the session, a pellet was automatically delivered every 8 min for 24 h. Water was available from a spout in the middle of the wall opposite the levers and food trough. The chambers were contained inside ventilated, sound attenuating cubicles, each with a 15-W light providing the same 12:12 cycle as the vivarium. Cue lights, real time records of pellet delivery and lever presses were controlled by Med-PC software (MED Associates Inc., St. Albans, VT). During all operant sessions, mice lived in the chambers with the exception of 30 min in the middle of each day when they were removed to a holding cage without food while the chambers were cleaned and serviced. Over the course of the next 3–5 days, mice received food contingent upon completion of a fixed ratio schedule; the ratio was increased (FR1 ! FR2 ! FR3) when mice showed sufficient lever pressing to maintain normal food intake. The mice were then studied at FR3 for another 7 days to establish baseline intakes. This phase of the experiment was conducted to establish whether mice would work for food; all mice performed well and were then moved to the PR schedule. Thereafter, because only four chambers were available, mice were run in squads each containing at least one mouse of each genotype.
2.2.
Progressive ratio sessions
On the PR1 schedule, the required responses for each successive pellet within a defined meal increased arithmetically (1, 2, 3, . . ., n). Each 23.5 h session began with the left cue and house lights on. Pressing could begin immediately and there was no additional cue for pellet delivery (i.e. no time out period). Whenever there was a 3-min stretch of no responding, the program reset. Resetting returned the lever press requirement to 1 and the animal was able to initiate a new period of pressing whenever desired. This reset criterion was studied for 12–13 days after which time mice were given free food in their home cage for 9–14 days. They were then returned to the chambers for another 12–13 days, but with a reset criterion of 20-min. In this situation, each session began with the cue and house lights off for 1 min. After the 1 min, the cue light was illuminated and mice could perform their PR run on this lever. During pellet delivery the house light came on briefly and was otherwise off. Whenever 20 min elapsed with no responding, the program reset. At the 20-min reset marker, a new run was signaled to the animal by a 1 min time out without cue light illumination. For both reset conditions, the number of responses, total pellets obtained, number of resets (defining meals), and average break point (defining meal size) was computed for
peptides 27 (2006) 2829–2835
2831
Table 1 – Mean daily food intake (g) during phases of experiment Phase FR3 PR 3-min reset PR 20-min reset
Wild type (n = 5)
Heterozygous (n = 5)
Knockout (n = 5)
4.2 0.1 4.0 0.1 4.2 0.1
4.6 0.1y 4.5 0.1 * 4.7 0.1 *
5.6 0.1 ** 5.2 0.1 ** 5.7 0.1 **
Values in table are the mean S.E. **Intake greater than HET and WT mice within phase ( p < 0.001). *Intake greater than WT within phase ( p < 0.001). yIntake greater than WT within phase ( p < 0.05).
each subject and session. Body weights were recorded at the beginning, middle, and end of each block of PR testing.
2.3.
Data analysis
The individual mean for total intake per day in each reset condition was calculated using 12 consecutive days of stable performance with correction for food spillage. Food spillage was counted as the number of pellets found beneath the grill floor, and this amount was subtracted from the daily intake. ANOVAs and Student t tests were used to analyze data and genotype was used as the grouping variable. The significance criterion was set at p < 0.05.
3.
Results
The mean intakes of food for each genotype in each phase of the experiment are shown in Table 1. In the FR3 condition, MC4RKO mice ate significantly more than the other two genotypes ( p < 0.001). This pattern was also present during the two PR phases, but in these phases the heterozygotes also showed a modest hyperphagia. The KO mice initially were obese, weighing 50% more than wild types and because of their sustained hyperphagia, their elevated weight was maintained throughout the experiment (Fig. 1 and Table 1). The average weight of the heterozygotes was slightly higher than the wild types. The mean responses per feeding bout from the PR phases are shown in Fig. 2. Also, shown above each bar is the mean number of pellets mice received before a breakpoint in responding. Breakpoints were lower during the 3 min than
Fig. 1 – Weights of mice during experimental phases. Group means are represented. KO > HET > WT, p < 0.001.
Fig. 2 – Mean responses per feeding bout for WT, HET and MC4RKO mice. Bars represent the mean W S.E. responses emitted per feeding bout including presses that did not directly result in pellet delivery. Numbers above each bar represent the mean number of pellets earned before a breakpoint: (A) shows responses in the 3-min reset criterion and (B) shows responses in the 20-min reset criterion. In (A) *MC4RKO mice had the highest responses per bout in comparison to WT ( p < 0.001) and HET ( p < 0.05). #HET and WT mice significantly differed ( p < 0.001) from each other. In (B) all genotypes significantly differed from each other (*KO > #HET > WT; p < 0.001).
the 20 min reset, and showed a significant effect of genotype with the highest breakpoints in the MC4RKO mice. On average, breakpoints were about three-fold higher in the 20 min compared with the 3 min condition, and were about two-fold higher in KO compared with wild type mice. The mean number of feeding bouts initiated during each PR phase is shown in Fig. 3. Mice took significantly fewer feeding bouts under the 20 min compared with the 3 min criterion, and KOs took slightly fewer feeding bouts than wild type. The behavior under the 3 min reset was essentially snacking, with about 40 feeding episodes per 23.5 h. The sum of the daily breakpoints across resets yields the total number of responses emitted per day, and these data are
2832
peptides 27 (2006) 2829–2835
Fig. 3 – Mean W S.E. number of snacks or small feeding bouts initiated by the mice under both PR1 conditions as denoted on the y-axis. Graph A represents data from the 3min reset condition. *Number of snacks were greater than HET and KO groups ( p < 0.001). Graph B represents data from the 20-min reset condition. *Number of snacks were greater than HET ( p < 0.05) and KO ( p < 0.001). #Number of snacks were greater than KO ( p < 0.05).
shown in Fig. 4. Total responses were greater at the 20 min reset than at the 3 min reset, and the MC4RKO mice emitted about two-fold more responses that the wild types, with the heterozygotes intermediate. The average price per pellet in the different reset conditions is shown in Table 2. KO and HET mice paid a higher average price per pellet WT and this effect was more pronounced in the 20-min reset phase. Fig. 5 shows 1 day of representative data for three mice, one from each genotype. Since mice were run in different squads, the start time (denoted as 0 on the x-axis) was approximately the same. These data show that more food was consumed at night compared with the daytime in all genotypes, although this was not formally analyzed.
4.
Discussion
To our knowledge this is the first report using a PR schedule in a closed economy to study meal patterns. PR schedules have been used predominantly in short sessions where breakpoints end the session, but in our case they defined elective meal
Fig. 4 – Twelve days average of total responses made per daily session. Graph A shows data from the 3-min reset criterion condition. *KO mice had a higher amount of responses daily than HET ( p < 0.01) and WT ( p < 0.001). # HET mice pressed more than WT ( p < 0.001). Graph B shows data from the 20-min reset criterion condition. *KO mice had a higher number of responses daily than HET and WT ( p < 0.001). #HET mice pressed more than WT ( p < 0.001) during this condition as well.
termination and so our interpretation of the breakpoint was most relevant to the assessment of the subjects’ ability to sense fullness and end a discrete bout of feeding. Under these conditions, the PR schedule may emulate depletion of a patch of finite resources, and the breakpoint represents a decision to move to a new patch. Additionally, PR schedules are thought to be particularly sensitive measures of motivation because they reveal the maximum price an individual is willing to pay for a fixed reinforcer [8,20]. The main purpose of the present study was to test a hypothesis that MC4RKO mice may be less motivated to work for food based on observations from a previous study [24]. In that study mice had a two lever contingency for pellet delivery; they were required to emit a certain number of responses (the procurement cost) on the left lever in order to for the right (pellet delivery) lever to be activated. At that point, a small fixed number of presses (5 or 10, the consumatory cost) on the
peptides 27 (2006) 2829–2835
Table 2 – Price (presses/number of pellets) averaged across 12 days of PR1 phases Phase PR 3-min reset PR 20-min resety
Wild type (n = 5)
Heterozygous (n = 5)
Knockout (n = 5)
3.8 0.1 5.6 0.1
5.2 0.1 ** 6.1 0.2 **
5.0 0.1 *** 7.7 0.2 ***
Values in table are the mean S.E. ***Price greater than HET and WT mice within phase ( p < 0.001). **Price greater than WT within phase ( p < 0.001). yOverall price greater than 3 min reset phase.
right lever delivered single pellets. However, after 10 min of no pressing on the right lever, the program de-activated this lever and in order to again obtain food again the animal had to run off a new procurement cost on the left lever. In both WT and MC4RKO mice, the number of meals per day declined but their size increased as the procurement cost rose, showing that the animals avoided work by changing their meal strategy. At the lowest cost on both levers, mice emitted 1000 presses per day and intakes were comparable for WT (3.6 g) and MC4RKO (4.0 g), and as a result the MC4RKO mice lost weight.
Fig. 5 – Cumulative responses of individual mice across a daily session; the actual start time for each session is indicated in the legend. Graph A shows cumulative responses for the 3-min reset condition. Graph B shows cumulative responses for the 20-min reset condition. Stepwise response increments are steeper during the 20-min reset condition for these three mice.
2833
In the present experiment, the results were strikingly different. First, in the training phase using an FR3 (consumatory cost only), MC4RKO mice were hyperphagic relative to WT. Critical differences from our previous study include the fact there was only one lever, and no ‘penalty’ for time without activity. These and other differences require further analysis. Our present results during the PR phases are inconsistent with our hypothesis and actually suggest the reverse. We expected that mice would adopt a strategy of small meals and frequent switching because the passage of time (reset criterion) was the only ‘‘cost’’ associated with initiating a new meal. This was especially evident at the lower (3 min) reset time. In part, we expected this strategy because previous studies using shorter sessions found differences between performance of ad libitum fed rodents and performance of rodents kept at 80% of their free feeding weight [14,17]. Our mice were not food deprived in this paradigm and thus determined their own pattern of feeding bouts. The low number of responses per feeding bout and frequent initiation of small bouts we observed during the 3min reset could in part also have been due to the lack of cue signaling. This may have confused the mice, which is why we added a cue and time out structure for the 20 min reset. Thus, the higher number of responses per feeding bout seen for all mice at this reset could have been due to the time, the cueing, or both. Future studies will be needed to examine in more detail the parameters of this protocol. Contrary to our hypothesis, KO mice did not engage in small, cheap meals but instead worked on average more per pellet than the other genotypes (Table 2). As proposed earlier as a possible outcome, we did see compensation by MC4RKO mice; their intermeal intervals were longer in the 3-min condition versus the 20 min condition (Fig. 5). Therefore, KO mice achieved hyperphagia under our PR1 schedules by having significantly higher meal sizes, as defined by number of responses per feeding bout, and lower meal frequencies, as defined by number of small feeding bouts, than the other genotypes. Breakpoints were higher for MC4RKO mice in the 20-min reset condition in comparison to the 3-min reset condition. At an average of 111 47 presses before taking a break, KO mice received about 14 pellets (0.28 g) per defined meal. During this period KO mice were initiating feeding bouts around 22 0.6 times per day. At the highest number of responses emitted per feeding bout of 158 presses, KO mice show an estimated breakpoint of about 17 pellets (0.34 g) in that discrete period of pressing. There was a reduction in the number of small feeding bouts seen for all mice to an average of 22–25 bouts per day during the 20-min reset condition. This shows that time available for feeding plays an important role in feeding strategy and in conjunction with the data on responses emitted (Fig. 4) suggest that the MC4R plays no role in adjusting effort to time constraints under these conditions. Due to the hyperphagic phenotype, it could be said that the KO mice were internally driven to maximize food intake, especially in the 20 min reset condition. Further evidence shows that MC4RKO mice paid a higher price per pellet than WT mice (Table 2) and thus can be considered more motivated than WT mice.
2834
peptides 27 (2006) 2829–2835
We examined the data for genotype-specific 24-h feeding patterns by analyzing records from three representative mice (Fig. 5). The three mice showed different profiles under the 3min versus the 20-min criterion. The representative WT nearly ceases feeding in the early morning hours of both reset conditions, consistent with previous work in mice [1]. The time constraint imposed on mice in the 20 min reset condition resulted in an increased rate of pressing for all genotypes. The KO and HET mice continue to lever press and eat well into the morning hours (Fig. 5). In an early study, Anliker and Mayer [1] found that ob/ob mice on an FR 25 schedule had very high response rate. ob/ob mice were hyperphagic on this FR schedule and showed no cyclical day/night pattern of eating. A similar consistent eating pattern characteristic of hyperphagia was also seen in the MC4RKO mouse in the present study (Fig. 5). Collier and colleagues [7] have shown that under two lever FR schedules in which work can be minimized by taking fewer larger meals, rats and many other species adopt exactly this strategy. It is thus not surprising that the converse of this pattern can be found under PR conditions in which work is avoidable by initiating small feeding bouts. We observed no decrease in weight when mice were in the chambers pressing for food. This suggests the effort of lever pressing did not contribute a significant metabolic demand. In past reports, hyperphagia in MC4RKO mice tends to be highly influenced by the housing environment [5,16,24]. It has been established that KO mice are missing an important receptor for the control of food intake but we still do not fully know what effect this has on postingestive feedback. The HET and WT mice may have more accurate negative feedback to end meals in comparison to KO mice and this may be most evident when mice have to work for food. HET and WT mice ate less than KOs but still maintained their body weights overall; this could be due to a number of external and/or internal factors. In summary, the deletion of the MC4R leads to increased food intake in mice [5,15,16,21,23]. Our previous experiment [24] did not find MC4RKO mice to be hyperphagic in the chamber, however when tested in a 24-h PR schedule MC4RKOs did exhibit significant and showed evidence of planning feeding bouts to maximize the amount of effort that contributed to feeding. The present experiment also introduces a novel protocol to study meal patterns: a closed economy PR protocol. We expected that this protocol would be a particularly stringent test of large meal phenotypes, and requires further validation using other protocols including for example chronic treatment with pharmacological agents that alter food intake. Through this study and others it is evident the MC4R gene positively influences descending information regarding feeding related behaviors. Though the MC4R is located in brain areas ascribed to be involved in motivation [3,19], the involvement of the MC4R seems to have the most significant role in the quantitative aspects of meal taking, like how much is eaten and how many calories to expend to gain access to food. This PR experiment has successful applied parameters of behavioral economics to MC4RKO feeding. Such an approach is useful tool in discovering how subjects adjust meal taking to different simulated environments.
Acknowledgements The MC4RKO mice were generously provided by Millennium Pharmaceuticals and this work was supported in part by NIHDK57080 (CHL).
references
[1] Anliker J, Mayer J. An operant conditioning technique for studying feeding-fasting patterns in normal and obese mice. J Appl Physiol 1956;8:667–70. [2] Balthasar N, Dalgaard LT, Lee CE, Yu J, Funahashi H, Williams T, et al. Divergence of melanocortin pathways in the control of food intake and energy expenditure. Cell 2005;123:493–505. [3] Berthoud HR. Multiple neural systems controlling food intake and body weight. Neurosci Biobehav Rev 2005;26:393–428. [4] Butler AA, Cone RD. Knockout studies defining different roles for melanocortin receptors in energy homeostasis. Ann NY Acad Sci 2003;994:240–5. [5] Butler AA, Marks DL, Fan W, Kuhn CM, Bartolome M, Cone RD. Melanocortin-4 receptor is required for acute homeostatic responses to increased dietary fat. Nat Neurosci 2001;4:605–11. [6] Chen AS, Metzger JM, Trumbauer ME, Guan X, Yu H, Frazier EG, et al. Role of the melanocortin-4 receptor in metabolic rate and food intake in mice. Transgenic Res 2000;9:145–54. [7] Collier GH, Johnson DF, Hill WL, Kaufman LW. The economics of the law of effect. J Exp Anal Behav 1986;46:113–36. [8] El-Ghundi M, O’Dowd BF, Erclik M, George SR. Attenuation of sucrose reinforcement in dopamine D1 receptor deficient mice. Eur J Neurosci 2003;17:851–62. [9] Farooqi IS, Keogh JM, Yeo GS, Lank EJ, Cheetham T, O’Rahilly S. Clinical spectrum of obesity and mutations in the melanocortin-4 receptor gene. N Engl J Med 2003;348:1096–103. [10] Findley JD. Preference and switching under concurrent scheduling. J Exp Anal Behav 1958;1:123–44. [11] Hadley ME, Haskell-Luevano C. The proopiomelanocortin system. Ann NY Acad Sci 1999;885:1–21. [12] Harrold JA, Widdowson PS, Williams G. Altered energy balance causes selective changes in melanocortin-4 (MC4R), but not melanocortin-3 (MC3-R), receptors in specific hypothalamic regions: further evidence that activation of MC4-R is a physiological inhibitor of feeding. Diabetes 1999;48:267–71. [13] Hodos W. Progressive ratio as a measure of reward strength. Science 1961;134:943–4. [14] Hodos W, Kalman G. Effects of increment size and reinforcer volume on progressive ration performance. J Exp Anal Behav 1963;6:387–92. [15] Huszar D, Lynch CA, Fairchild-Huntress V, Dunmore JH, Fang Q, Berkemeier LR, et al. Targeted disruption of the melanocortin-4 receptor results in obesity in mice. Cell 1997;88:131–41. [16] Irani BG, Xiang Z, Moore MC, Mandel RJ, Haskell-Luevano C. Voluntary exercise delays monogenetic obesity and overcomes reproductive dysfunction of the melanocortin-4 receptor knockout mouse. Biochem Biophys Res Commun 2005;326:638–44. [17] Jewett DC, Cleary J, Levine AS, Schaal DW, Thompson T. Effects of neuropeptide Y, insulin, 2-deoxyglucose, and food deprivation on food-motivated behavior. Psychopharmacology 1995;120:267–71.
peptides 27 (2006) 2829–2835
[18] Low MJ, Hayward MD, Appleyard SM, Rubinstein M. Statedependent modulation of feeding behavior by proopiomelanocortin-derived beta-endorphin. Ann NY Acad Sci 2003;994:192–201. [19] Mountjoy KG, Mortrud MT, Low MJ, Simerly RB, Cone RD. Localization of the melanocortin-4 receptor (MC4-R) in neuroendocrine and autonomic control circuits in the brain. Mol Endocrinol 1994;8:1298–308. [20] Richardson NR, Roberts DCS. Progressive ratio schedules in drug self-administration studies in rats: a method to evaluate reinforcing efficacy. J Neurosci Meth 1996;66:1–11. [21] Samama P, Rumennik L, Grippo JF. The melanocortin receptor MCR4 controls fat consumption. Regul Pept 2003;113:85–8.
2835
[22] Stafford D, LeSage MG, Glowa JR. Progressive-ratio schedules of drug delivery in the analysis of drug selfadministration: a review. Psychopharmacology 1998;139:169–82. [23] Ste Marie L, Miura GI, Marsh DJ, Yagaloff K, Palmiter RD. A metabolic defect promotes obesity in mice lacking melanocortin-4 receptors. Proc Natl Acad Sci USA 2000;97:12339–44. [24] Vaughan CH, Moore MC, Haskell-Luevano C, Rowland NE. Meal patterns and foraging in melanocortin receptor knockout mice. Physiol Behav 2005;84:129–33. [25] Vergoni AV, Schioth HB, Bertolini A. Melanocortins and feeding behavior. Biomed Pharmacother 2000;54: 129–34.