Behavioral Tests of the Insulin-Cholinergic-Dopamine Link in Nucleus Accumbens and Inhibition by High Fat-High Sugar Diet in Male and Female Rats

Abstract

It was previously shown in striatal slices obtained from male rats that insulin excites cholinergic interneurons and increases dopamine (DA) release via α4β2 nicotinic receptors on DA terminals. The effect of insulin on DA release was blocked either by maintaining rats on a high sugar-high fat (HS-HF) diet that induced hyperinsulinemia and nucleus accumbens (NAc) insulin receptor insensitivity, or applying the α4β2 antagonist DHβE. In vivo, NAc shell insulin inactivation decreased a glucose lick microstructure parameter indicative of hedonic impact in male and female rats, and prevented flavor-nutrient learning, tested only in males. The HS-HF diet decreased hedonic impact in males but not females, and prevented flavor-nutrient learning, tested only in males. The present study extends testing to more fully assess the translation of brain slice results to the behaving rat. Insulin inactivation by antibody microinjection in NAc shell was found to decrease the number of lick bursts emitted and average lick burst size, measures of incentive motivation and hedonic impact respectively, for a wide range of glucose concentrations in male and female rats. In contrast, the HS-HF diet decreased these lick parameters in males but not females. Follow-up two-bottle choice tests for 10% versus 40% glucose showed decreased intake of both concentrations by males but increased intake of 40% glucose by females. In a further set of experiments, it was predicted that α4β2 receptor blockade would induce the same behavioral effects as insulin inactivation. In females, DHβE microinjection in NAc shell decreased both lick parameters for glucose as predicted, but in males only the number of lick bursts emitted was decreased. DHβE also decreased the number of lick bursts emitted for saccharin by females but not males. Finally, DHβE microinjection in NAc shell decreased flavor-nutrient learning in both sexes. The few discrepancies seen with regard to the hypothesized insulin-nicotinic-dopaminergic regulation of behavioral responses to nutritive sweetener, and its inhibition by HS-HF diet, are discussed with reference to sex differences in DA dynamics, female resistance to diet-induced metabolic morbidities, and extra-striatal cholinergic inputs to NAc.

1. Introduction

Insulin of peripheral origin is an established adiposity signal with a medial hypothalamic site of anorexigenic action (1). Several extra-hypothalamic targets of peripheral insulin have also been identified, including nucleus accumbens (NAc) where insulin facilitates the release of dopamine and glutamate, and regulates appetitive motivation, the hedonic response to glucose, and flavor-nutrient learning (2–7).

It was previously shown that supplementation of a standard chow diet with continuous access to a high fat-high sugar (HF-HS) liquid (chocolate Ensure) induces NAc insulin receptor insensitivity, impairs flavor-nutrient learning, and blunts the hedonic impact of glucose, but not saccharin, as measured by lick microstructure analysis (6,8). The plausible involvement of impaired NAc insulin signaling in the behavioral effects of HF-HS supplementation was supported by the strong positive correlation between consummatory anhedonia and diet-induced hyperinsulinemia (8), and duplication of the HF-HS effects on hedonic reactivity and flavor-nutrient learning by NAc insulin inactivation (2,5).

The current study begins by addressing a question that arises from the previous use of a single glucose concentration in the lick microstructure experiments. Just as single dose pharmacology engenders the risk of misleading conclusions, lick microstructure testing to probe hedonic impact and motivation with a single glucose concentration may not provide an accurate portrait of HF-HS and NAc insulin effects. Consequently, the first experiment of the present study used lick microstructure analysis with glucose concentration-response testing in chow-fed rats receiving NAc microinjection of insulin-inactivating antibody and rats whose chow diet was supplemented with continuous access to HF-HS liquid.

Results obtained in this experiment suggested that consumption of a HF-HS diet may lead to a right-shift in glucose concentration preference. This could have translational implications including exacerbation of body weight gain and metabolic morbidities. Consequently, in a second experiment, relative preference for 10% versus 40% glucose in a two-bottle access protocol was tested in rats maintained on lab chow versus HF-HS liquid supplementation.

Cholinergic interneurons (ChIs) in NAc, which stimulate dopamine (DA) release via nicotinic receptors (nAChRs) on DA axons (9–12), show high expression of insulin receptors and respond to insulin with increased excitability followed by a marked increase in extracellular DA (2,13). The NAc DA response to insulin is blocked by the α4β2 nicotinic antagonist, DHβE, and is lost in striatal slices obtained from male rats maintained on the HF-HS diet (2). Given that lick microstructure parameters reflective of motivation and reward, and flavor-nutrient learning are dependent on DA receptor signaling in NAc (14–18), the third and fourth experiments of the present study test the prediction that blockade of α4β2 nAChRs in NAc shell will have the same behavioral effects seen in response to insulin inactivation and dietary supplementation with a HF-HS liquid.

An overarching question addressed here concerns possible sex differences in treatment effects. Most of the key previous experiments were limited to male rats, but behavioral results obtained with liquid HF-HS supplementation as well as a solid HF-HS diet indicated differences between male and female rats. Specifically, HF-HS liquid supplementation, which induced anhedonia in males, did not decrease the hedonic response to glucose in females (5), and while a negative correlation was seen between peripheral insulin level and hedonic impact (average lick burst size) in males, a positive correlation was seen between peripheral insulin level and motivation to consume glucose (number of lick bursts emitted) in females (8). Further, a solid HF-HS diet that induced anhedonia in males, produced the opposite effect, enhancing the hedonic response to glucose in females (5). Consequently, cohorts of male and female rats were tested throughout the present study. A detailed schematic depicting the relationship between insulin receptors, cholinergic interneurons, DA neurons and DA release in NAc, as they relate to hypotheses tested in the present study, is available in a recent review (13).

2. Materials and methods

2.1. Animals

Male and female Sprague–Dawley rats were purchased from Taconic Farms (Germantown, NY) at 10–12 weeks of age. Rats were housed individually in plastic cages with bedding and free access to water. They were maintained on a 12 h light/dark cycle (lights on at 6 am) and allowed at least three days acclimation to vivarium housing prior to initiation of any experimentation. All experimental procedures were approved by the New York University Grossman School of Medicine Institutional Animal Care and Use Committee and were performed in accordance with the “Principles of Laboratory Animal Care” (NIH publication number 85–23). All efforts were made to minimize animal suffering and to reduce the number of animals used.

2.2. Diets

All rats had free access to standard lab pellets (Rodent Diet #5001, Lab Diet, St. Louis, MO) and tap water in the home cage. None of the experiments described included a period of fasting. One of the three groups of rats included in the lick microstructure testing of multiple glucose concentrations had continuous home cage access to Milk Chocolate Ensure in addition to chow. The ratio of % kcal from fat:carb:protein for this HF-HS liquid supplement is 4:4:2 with 25% sugars (Abbott, Columbus, OH). Rats were maintained on their assigned diets for six weeks prior to initiation of experimental testing. We have previously shown that six weeks of Ensure supplementation increases circulating insulin and leptin levels relative to control in both male and female rats (2,8).

2.3. Surgery and intracerebral microinjection

Two of the three groups of rats in the glucose concentration-response experiment, and all subjects in the DHβE microinjection experiments, were anesthetized with ketamine (100 mg/kg, i.p.) and xylazine (10 mg/kg, i.p.), and implanted with chronically indwelling guide cannulas (P1 Technologies, Roanoke, VA). Two cannulas (26 ga) were placed bilaterally 2.0 mm dorsal to the injection sites in the NAc shell (1.6 mm anterior to bregma; 2.1 mm lateral to the sagittal suture, tips angled 8° toward the midline, 5.8 mm ventral to skull surface). Patency was maintained with occlusion stylets. The cannulas and four mounting screws were then permanently secured to the skull with dental acrylic. Postsurgical analgesia was achieved by administration of Carprofen (5.0 mg/kg, s.c.) immediately post-surgery and twenty-four hours later.

After one week recovery from surgery, rats were habituated on five occasions to 2–3 min light restraint on a foam cushion and received mock microinjections in preparation for subsequent experimental microinjection treatments. Immediately prior to behavioral test sessions, rats in the glucose concentration-response experiment received bilateral microinjection of either undiluted InsAb (concentration 1μg/μl; low endotoxin, azide-free; AB46707, Abcam, Cambridge MA) as validated previously (2,6) or the undiluted control IgG1 (low endotoxin, azide-free; AB18437, Abcam, Cambridge MA). Rats in the experiment testing for nicotinic cholinergic regulation received bilateral injection of either DHβE (2 μg and 10 μg; Tocris, Minneapolis MN) or saline vehicle. Rats in the conditioned flavor preference experiment received bilateral injection of either DHβE (10 μg) or saline vehicle immediately prior to conditioning sessions. For microinjections, two 30 cm lengths of PE-50 tubing were attached on one side to 5-μL Hamilton syringes filled with distilled water, and on the other side to 31-gauge injector cannulas extending 2.0 mm beyond the implanted guides. Rats were placed on a cushion and their stylets were removed. Injectors were inserted and microinjections were made at rate of 0.05 μL every 10 seconds over a period of 100 s. Injector cannulas were left in place for another minute before being removed, at which point stylets were replaced and the rat was placed in the behavioral test chamber where an experimental session was initiated.

2.4. Lick microstructure

In preparation for lick microstructure testing, rats were habituated to the test chamber on 3–4 occasions and given the opportunity to develop reliable licking of 6% glucose. A two-minute bout of licking was considered criterion for lick acquisition and terminated the session which would otherwise continue for 30 min. This was followed by three 30 min pretest sessions, on consecutive days, in which the total number of licks emitted was used to form matched groups of rats for subsequent experimental testing. In order to eliminate the possibility of interference from sex-related chemosensory or ultrasonic stimuli in the test room or chambers, males and females were run as separate cohorts at different times and therefore treated as separate experiments. In the glucose concentration-response experiment one group of each sex received NAc microinjection of InsAb (males: n=13; females: n=10) and the other the control IgG (males: n=10; females: n=9). A third group of rats of each sex (males: n=11; females: n=7) that had been maintained with Ensure supplementation for six weeks underwent the same sequence of habituation and pretesting, with no additional treatment prior to test sessions. Table 1 indicates the timeline and procedures of these experiments.

Table 1:

Timeline and procedures for determining treatment effects on glucose lick microstructure and follow-up testing of glucose concentration preference.

	Week 1	Week 2	Week 3	Week 4 & 5	Week 6	Week 7	Week 8	Week 9
Group (IgG)		Surgery	Recovery	Habituation Lick Acquisition	3 Pretests	4 Tests		2-Bottle Acquisition 2-Bottle Testing
Group (InsAb)		Surgery	Recovery	Habituation Lick Acquisition	3 Pretests	4 Tests
Group (Ensure)	Start Ensure			Habituation Lick Acquisition	3 Pretests	4 Tests		2-Bottle Acquisition 2-Bottle Testing

In the DHβE experiments all rats were tested under the two doses of DHβE and vehicle. In one DHβE experiment for each sex the available fluid was 6.1% glucose. In a second experiment the available fluid was 0.25% saccharin (males/glucose: n=11; males/saccharin: n=12; females/glucose: n=11; females/saccharin: n=10).

Testing was conducted in Med Associates (Georgia, VT) operant chambers with a contact lickometer using Med PC and custom software. Each rat was tested in four 30 min sessions spaced 48 h apart. In the glucose concentration-response experiment a different concentration of unflavored glucose, 2%, 10%, 20% or 40%, was available in each session with half the rats in each treatment group (InsAb, IgG, HF-HS) receiving concentrations in ascending order and half in descending order. In the DHβE experiment the fluid available in test sessions was either 6% glucose or 0.25% saccharin flavored with 0.05% grape Kool-Aid (Kraft Foods, Northfield, IL). Grape flavor was used to match, as far as possible, taste properties of the glucose and saccharin solutions, and to obtain results applicable to planning of the conditioned flavor preference experiment to follow (see below). Saline vehicle was administered in the first and fourth test sessions and results were averaged to yield a control value. In the second and third test sessions half the rats were tested with the 2-μg dose and 10-μg dose, respectively, and half in the reverse order. The lick parameters recorded included total number of licks per session, the total number of lick bursts, which are groups of licks separated by an inter-lick interval of > 1 sec, and the average lick burst size, which is the number of licks in a burst, where only bursts of > 3 licks are counted. The number of bursts emitted per unit time is considered a measure of motivation to consume while burst size is considered a measure of hedonic impact (19–26). All behavioral data were automatically collected by computer.

2.5. Glucose concentration preference

Rats that had received control microinjections and HF-HS supplementation in the lick microstructure experiments continued their assigned diets (i.e., chow or chow + Ensure). For testing, each rat was placed in a plastic cage containing two bottles: one on the left and one on the right. Cages did not contain bedding and food was unavailable. One bottle contained 40 ml of 10% glucose and the other contained 40 ml of 40% glucose. Side placement of the two concentrations was alternated across four sessions on consecutive days. The first two sessions were considered “acclimation/acquisition”. The volume of each solution consumed in the third and fourth sessions was averaged for each rat. Results were expressed as absolute volume consumed and the ratio of 40% glucose to 10% glucose consumed.

2.6. Conditioned flavor preference

To test predictions that DHβE will duplicate the effect of InAb, males (n=9) and females (n=10) were tested for acquisition of flavor preference using 6.1% glucose + 0.2% saccharin as the sweetener. The reason for use of 6.1% glucose in both the DHβE lick microstructure and flavor-nutrient conditioning experiments was to match the glucose solution used in the original study of insulin antibody effects on glucose lick microstructure (5) and the glucose solution used in the previous studies of HF-HS diet and insulin antibody effects on acquisition of flavor-nutrient learning (6).

In preparation for conditioning, each rat underwent four 30 min sessions in a clear plastic cage without bedding that contained a left-side and right-side bottle. Both bottles were filled with unflavored 6.1% glucose. Each of these sessions was preceded by a mock microinjection treatment consisting of placement on a foam cushion and having cannula stylets removed and replaced. These sessions were followed by eight conditioning sessions, held on consecutive days except for weekends. In each 30 min session there were two bottles filled with flavored 6.1% glucose + 0.2% saccharin. On alternate days, sessions were preceded by bilateral microinjection of 10 μg DHβE or vehicle. On vehicle microinjection days both bottles of sweetened fluid were flavored with 0.05% grape Kool-Aid (Kraft Foods, Northfield, IL) for half of the rats and 0.05% cherry Kool-Aid for the remaining half. On DHβE microinjection days the flavors were reversed. The starting volume of solution in each of the two bottles was limited to 5 ml on each side (with additional 6 ml for dead space) based on the total 30 min consumption of DHβE-treated subjects in the lick microstructure experiments. The purpose of this volume limit was to equalize intake during vehicle and DHβE conditioning sessions.

Twenty-four hours after the final conditioning test rats underwent two 30 min flavor preference tests on consecutive days. One bottle contained 50 ml of grape flavored 0.2% saccharin and the other 50 ml of cherry flavored 0.2% saccharin. Side locations of the two flavors were reversed between the two tests and volume of each flavor consumed was averaged to yield test results for each rat.

2.7. Histology

Rats were briefly exposed to CO₂ and decapitated by guillotine. Brains were removed and fixed in 10 % buffered formalin for 48 h. Frozen 40 μm coronal sections were cut on a Reichert-Jung 2800 cryostat, thaw-mounted on gelatin-coated slides and stained with cresyl violet. Microinjection sites were determined by visual inspection under an Olympus SZ40 microscope.

2.8. Data Analysis

Results for each lick microstructure parameter in the glucose concentration-response experiment were analyzed separately for each sex by two-way ANOVA with repeated measures on the factor of glucose concentration, followed by Fisher’s protected t-tests where appropriate and of interest.

Results of the two-bottle choice experiment were also analyzed for each sex by two-way ANOVA with repeated measures on the factor of glucose concentration, followed by Fisher’s protected t-tests where appropriate and of interest. In addition, results for each rat were expressed as the ratio of volumes consumed (i.e., 40%:10%) and the comparison between diet groups of each sex was analyzed by t-test.

Results for each lick microstructure parameter in the DHβE experiments were analyzed for each sex by one-way repeated measures ANOVA followed by Fisher’s protected t-test where appropriate.

For conditioned flavor preference the difference between volumes of the DHβE- and vehicle-paired flavors consumed were first analyzed by t-test for correlated samples. This was followed by calculation of the percentage decrease in consumption of the DHβE- relative to vehicle-paired flavor.

3. Results

3.1. Effect of InsAb and HF-HS diet on lick microstructure for glucose concentrations

The hypothesis tested in this experiment is that insulin inactivation and HF-HS diet both decrease glucose licking behavior and do so in a manner that is exclusively or preferentially expressed in the measure of hedonic impact, namely the average number of licks per burst. Mechanistically, this is based on the previous findings that HF-HS diet induced NAc insulin insensitivity in male rats, inhibited insulin-induced DA release, and NAc insulin inactivation and HF-HS diet both decreased the hedonic impact of single concentration glucose in male rats (2,5,6).

Males

Consumption of Ensure and chow

In the group of male rats whose treatment consisted of diet supplementation with continuous access to milk chocolate Ensure, the weekly mean intakes of Ensure and chow are shown in Figure 1. In week seven, when glucose concentration-response testing was conducted, the mean 24 h caloric intake from Ensure and chow were 64.2 and 73.8 kcal, respectively. The combined 24 h caloric intake of 138 kcal was slightly greater than the 128.5 kcal of a representative group of similar weight males in the laboratory that were maintained on chow exclusively (15).

Figure 1. Twenty-four-hour intake of milk chocolate Ensure and lab chow by male and female rats.

Twenty-four-hour intake of Ensure (top; mean ± SEM) and chow (bottom; mean ± SEM) were measured once per week in the groups of male and female rats for which HF-HS diet was the experimental treatment. Glucose concentration-response testing of lick microstructure was conducted during week seven (see Table 1). * For purposes of comparison, chow-fed male and female rats whose initial body weights were similar to rats depicted here consumed an average of 31.5 ± 1.8 and 23.9 ± 1.5 grams, respectively, in twenty-four hours (15).

Total number of licks:

Results for total number of licks emitted are displayed in Figure 2 (top panel). ANOVA that included the three treatment groups indicated significant main effects of treatment and glucose concentration (F(treatment)_2,32=43.3, p<.0001; F(concentration)_3,96=22.7, p<.0001). Follow-up ANOVAs on each combination of two groups indicated that total number of licks emitted was lower in the InsAb-treated (F(treatment)_1,22=65.1, p<.0001) and HF-HS-treated group (F(treatment)_1,20=39.3, p<.0001), relative to the IgG-treated group. However, the decreasing effect of InsAb treatment was greater than HF-HS (F(treatment)_1,22=12.7, p<.01).

Figure 2.

Effects of nucleus accumbens shell insulin antibody microinjection and high fat-high sugar liquid diet supplementation on microstructure of licking for glucose in male rats. Top: Total number of licks emitted (mean ± SEM) for 2%, 10%, 20% and 40% glucose solution in a series of four 30-min sessions. Effects of NAc insulin antibody (InsAb) microinjection (red; n=13), diet supplementation with continuous access to chocolate Ensure (green; n=11), and the control treatment of immunoglobulin G (IgG) microinjection in nucleus accumbens (blue; n=11) are compared. ** p at least <.01.

Middle: Number of lick bursts emitted (mean ± SEM) for 2%, 10%, 20% and 40% glucose solution to yield the total lick count depicted (Top). ** p at least <.0001.

Bottom: Average number of licks per burst (mean ± SEM) as depicted (Middle) for 2%, 10%, 20% and 40% glucose solution. ** p at least <.01.

Number of lick bursts:

Results for number of lick bursts emitted are displayed in Figure 2 (middle panel). ANOVA with the three treatment groups indicated significant main effects of treatment and glucose concentration (F(treatment)_2,32=24.3, p<.0001; F(concentration)_3,96=8.98, p<.0001). Follow-up ANOVAs indicated that the number of lick bursts emitted was lower in both the InsAb-treated (F(treatment)_1,20=38.11, p<.0001) and HF-HS-treated group (F(treatment)_1,20=25.1, p<.0001) relative to the IgG-treated group, with no difference between the HF-HS-treated and InsAb-treated groups (F(treatment)_1,22=2.61, p>.10).

Lick burst size:

Results for average number of licks per burst are displayed in Figure 2 (bottom panel). ANOVA with the three treatment groups indicated significant main effects of treatment and glucose concentration (F(treatment)_2,32=7.81, p<.01; F(concentration)_3,96=7.68, p<.001). Follow-up ANOVAs indicated that lick burst size was smaller in both the InsAb-treated (F(treatment)_1,22=11.67, p<.01) and HF-HS-treated group (F(treatment)_1,22=8.58, p<.01) relative to the IgG-treated group, with no difference between the HF-HS-treated and InsAb-treated groups (F(treatment)_1,22=0.79, p>.10).

Females

Consumption of Ensure and chow

In the group of female rats whose treatment consisted of diet supplementation with continuous access to milk chocolate Ensure, the weekly mean intakes of Ensure and chow are shown in Figure 1. In week seven, when glucose concentration-response testing was conducted, mean 24 h caloric intake from Ensure and chow were 56.7 and 51.4 kcal, respectively. The combined 24 h caloric intake of 108.1 kcal was slightly higher than 97.5 kcal of a representative group of similar weight females in the laboratory that were maintained on chow exclusively (8).

Total number of licks:

Results for total number of licks emitted are displayed in Figure 3 (top panel). ANOVA that included the three treatment groups indicated significant main effects of treatment and glucose concentration (F(treatment)_2,24=14.1, p<.0001; F(concentration)_3,72=56.4, p<.0001). Follow-up ANOVAs on each combination of two groups indicated that total number of licks emitted was lower in the InsAb-treated (F(treatment)_1,17=21.4, p<.0001) but not the HF-HS-treated group (F(treatment)_1,15=0.68, p>.10) relative to the IgG-treated group. In addition, the number licks emitted by the HF-HS-treated group was greater than that of the InsAb-treated group (F(treatment)_1,16=20.46, p<.0001).

Figure 3. Effects of nucleus accumbens shell insulin antibody microinjection and high fat-high sugar liquid diet supplementation on microstructure of licking for glucose in female rats.

Top: Total number of licks emitted (mean ± SEM) for 2%, 10%, 20% and 40% glucose solution in a series of four 30-min sessions. Effects of NAc insulin antibody (InsAb) microinjection (red; n=10), diet supplementation with continuous access to chocolate Ensure (green; n=8), and the control treatment of immunoglobulin G (IgG) microinjection in nucleus accumbens (blue; n=9) are compared. ** p at least <.0001.

Middle: Number of lick bursts emitted (mean ± SEM) for 2%, 10%, 20% and 40% glucose solution to yield the total lick count depicted (Top). ** p at least <.01.

Bottom: Average number of licks per burst (mean ± SEM) as depicted (Middle) for 2%, 10%, 20% and 40% glucose solution. * p <.05.

Number of lick bursts:

Results for total number of lick bursts emitted are displayed in Figure 3 (middle panel). ANOVA with the three treatment groups indicated significant main effects of treatment and glucose concentration (F(treatment)_2,24=16.82, p<.0001; F(concentration)_3,72=53.12, p<.0001). Follow-up ANOVAs indicated that the number of lick bursts emitted was lower in the InsAb-treated group (F(treatment)_1,17=15.76, p<.001) but not the HF-HS-treated group (F(treatment)_1,15=1.09, p>.10) relative to the IgG-treated group. In addition, the number of lick bursts emitted by the HF-HS-treated group was greater than that of the InsAb-treated group (F_1,16=51.92, p<.0001). Interaction between treatment and glucose concentration (F_,3,45=5.04, p<.01) indicated that a greater number of lick bursts were emitted for 20% glucose by the HF-HS than IgG group (t(45)=3.45, p<.01).

Lick burst size:

Results for average number of licks per burst are displayed in Figure 3 (bottom panel). ANOVA with the three treatment groups indicated significant main effects of treatment and glucose concentration (F(treatment)_2,24=4.2, p<.05; F(concentration)_3,72=9.28, p<.0001). Follow-up ANOVAs indicated that burst size was lower in the InsAb-treated group (F(treatment)_1,17=6.55, p<.05) but not the HF-HS-treated group (F(treatment)_1,19=1.94, p>.10) relative to the IgG-treated group. However, lick burst size was not significantly greater in the HF-HS group than the InsAb-treated group (F(treatment)_1,16=3.73, p=.071).

3.2. Two bottle choice for glucose concentrations

The hypothesis tested in this experiment, based on the HF-HS diet-induced decrease in glucose lick parameters in males but not females, is that diet supplementation with chocolate Ensure will shift relative preference from 10% to 40% glucose in male but not female rats.

Males

Results for absolute volumes consumed in two-bottle tests are displayed in Figure 4. ANOVA indicated a main effect of diet (F_1,19=52.8, p<.0001) wherein glucose intake was decreased in HF-HS supplemented relative to chow-fed males, and a main effect of glucose concentration with greater intake of 40% relative to 10% glucose (F_1,19=97.2, p<.000). When results were expressed as the ratio of 40% to 10% glucose consumed (Figure 5) no difference was seen between the diet groups (t(19)=0.94).

Figure 4. Effect of diet on consumption of 10% versus 40% glucose in a two-bottle choice test.

Left: Volumes (ml; mean ± SEM) consumed by male rats in a 30 min test session. Results shown for each rat maintained on lab chow or lab chow with HF-HS liquid (Ensure) supplementation are the mean of two sessions in which bottle placements were alternated between left and right sides of the test cage. ***p<.0001 Right: Volumes (ml; mean ± SEM) consumed by female rats in a 30 min test session. Results shown for each rat maintained on either lab chow or lab chow with HF-HS liquid (Ensure) supplementation are the mean of two sessions in which bottle placements were alternated between left and right sides of the test cage. **p<.001

Figure 5. Ratio of 40%:10% glucose consumed in diet groups.

Results depicted in Figure 4 expressed as the ratio of 40% to 10% glucose consumed by male and female rats in the chow-only versus chow with HF-HS liquid supplementation diet groups. *p<.05

Females

Results for absolute volumes consumed in two-bottle tests are displayed in Figure 4. ANOVA did not indicate a main effect of diet (F_1,13=3.04, p=.10) but did indicate greater consumption of 40% relative to 10% glucose (F_1,13=257.6, p<.0001). Interaction between diet and glucose concentration (F_1,13=45.08, p<.0001) followed by Fisher’s protected t-test indicated that HF-HS rats consumed a greater volume of 40% glucose than the chow-fed group (t(13)=7.4, p<.001). When results were expressed as the ratio of 40% to 10% glucose consumed (Figure 5), HF-HS supplementation was shown to increase the ratio (t(13)=2.17, p<.05).

3.3. Effect of DHβE on lick microstructure

The hypothesis tested in this experiment is that blockade of α4β2 nicotinic receptors will have the same effect as insulin inactivation. Based on results of the preceding experiment, the predicted outcome would be a decrease in all glucose lick microstructure parameters. Mechanistically, this is based on the finding that NAc insulin excites ChIs, which in turn increase DA release via α4β2 receptors (2). Further, based on previous findings (2,5,6), it is predicted that similar to InsAb, blockade of α4β2 nicotinic receptors will have no effect on licking for the non-nutritive sweetener, saccharin.

Males: Glucose

Number of lick bursts:

Results for the number of lick bursts emitted are displayed in Figure 6 (top, left). ANOVA indicated that the number of lick bursts emitted differed across treatments (F_2,20=6.05, p<.01) with the 10 μg dose having a decreasing effect relative to vehicle (t(20)=3.0, p<.01) and the 2 μg dose having no effect (t(20)=0.46).

Figure 6. Effect of DHβE microinjection in nucleus accumbens shell of male rats.

Results depicted are for number of lick bursts emitted and lick burst size for flavored 6.1% glucose (top) and 0.25% saccharin (bottom) in male rats. *p<.05, **p<.01

Lick burst size:

Results for the average number of licks per burst are displayed in Figure 6 (top, right). DHβE had no effect on lick burst size (F_2,20=0.18).

Males: Saccharin

Number of lick bursts:

Results for the number of lick bursts emitted are displayed in Figure 6 (bottom, left). For saccharin, the number of lick bursts emitted differed across treatments (F_2,22=6.06, p<.01) with the 2 μg dose having an increasing effect relative to vehicle (t(22)=2.08, p<.05) and the 10 μg dose having no effect (t(22)=1.44).

Lick burst size:

Results for the average number of licks per burst are displayed in Figure 6 (bottom, right). The average lick burst size differed across treatments (F_2,22=7.22, p<.01) with the 2 μg dose having an increasing effect relative to vehicle (t(22)=3.11, p<.01) and the 10 μg dose having no effect (t(22)=0.13).

Females: Glucose

Number of lick bursts:

Results for the number of lick bursts emitted are displayed in Figure 7 (top, left). For glucose, the number of lick bursts emitted differed across treatments (F_2,20=24.6, p<.0001) with the 2 μg dose (t(20)=3.4, p<.01) and 10 μg dose (t(20)=7.3, p<.001) having decreasing effects relative to vehicle.

Figure 7. Effect of DHβE microinjection in nucleus accumbens shell of female rats.

Results depicted are for number of lick bursts emitted and lick burst size for flavored 6.1% glucose (top) and 0.25% saccharin (bottom) in female rats. **p<.01, ***p<.001

Lick burst size:

Results for the average number of licks per burst are displayed in Figure 7 (top, right). The average lick burst size differed across treatments (F_2,20=18.75, p<.0001) with the 2 μg dose (t(20)=3.4, p<.01) and 10 μg dose (t(20)=7.3, p<.001) having decreasing effects relative to vehicle.

Females: Saccharin

Number of lick bursts:

Results for the number of lick bursts emitted are displayed in Figure 7 (bottom, left). For saccharin, the number of lick bursts emitted differed across treatments (F_2,18 =29.2, p<.001) with the 2 μg dose (t(18)=5.84, p<.001) and 10 μg dose (t(18)=7.22, p<.001) having decreasing effects relative to vehicle.

Lick burst size:

Results for the average number of licks per burst are displayed in Figure 7 (bottom, right). The average lick burst size did not differ across treatments (F_2,18 =2.2, p=.14).

3.4. Effect of DHβE on conditioned flavor preference

The hypothesis tested in this experiment is that blockade of α4β2 nicotinic receptors will have the same effect as insulin inactivation, namely inhibition of flavor-nutrient learning (2,6). Mechanistically, this is based on the finding that NAc insulin excites ChIs, which in turn increase DA release via α4β2 receptors (2).

Results for flavor preference conditioning are displayed in Figure 8. Male rats showed a modest but significant decrease in preference for the DHβE-paired flavor (t(16)=3.26, p<.01). Female rats showed a marked decrease in preference for the flavor associated with DHβE pretreatment, consuming approximately twice the volume of the vehicle-paired relative to the DHβE-paired flavored solution during testing (t(18)=7.2, p<.0001).

Figure 8. Effect of DHβE microinjection in nucleus accumbens shell on flavor-nutrient learning.

Conditioning was conducted with grape versus cherry flavored 6.1% glucose + 0.2% saccharin solution. Testing was conducted with two-bottle choice of cherry or grape flavored 0.2% saccharin solution. Results depicted are volumes of DHβE-paired flavor and vehicle-paired flavor consumed by male (left) and female (right) rats. Results are average of two tests with flavor locations switched between left and right. **p<.01, ***p<.001

4. Discussion

Peripheral insulin enters the brain by trans-endothelial transport (27,28), and there is a rapid increase in numerous brain regional insulin levels during/after a meal (29). We previously observed NAc insulin receptor phosphorylation and downstream Akt signaling immediately after completion of a seven-minute intragastric infusion or oral consumption of glucose (6). Further, supplementation of laboratory chow with continuous access to HF-HS liquid, which induced peripheral hyperinsulinemia and NAc insulin receptor insensitivity, was shown to induce behavioral changes in glucose lick microstructure and flavor-nutrient learning (6,8).

In the previous lick microstructure study with single concentration glucose (8) HF-HS-supplemented male rats displayed a decrease in number of lick bursts emitted, an indication of decreased incentive motivation, and lick burst size, an indication of decreased hedonic impact, across six weekly test sessions, though the effect on number of lick bursts was not significant in the final test session on week six (8). In contrast, female rats displayed only a decrease in number of lick bursts emitted across the six weekly test sessions, with the effect no longer significant on week six.

The present experiment, which began after rats had been on their assigned diet for six weeks, and probed a range of glucose concentrations, clarifies both the diet effect and the sex difference. In males, HF-HS decreased both the number of lick bursts emitted and the average lick burst size across glucose concentrations. In females, HF-HS failed to decrease either lick parameter. In fact, HF-HS supplementation tended to increase the number of lick bursts emitted, with the effect being statistically significant for the 20% glucose concentration. The sex difference could not be attributed to decreased home cage consumption of Ensure by females; in fact, average 24 h intake of Ensure throughout the experiment was greater in females than in males, at 0.26 ml versus 0.17 ml per gram of body weight.

The sex difference also cannot likely be attributed to different consequences of NAc insulin receptor insensitivity in as much as NAc insulin inactivation markedly decreased all lick parameters across glucose concentrations in both males and females. This result also clarifies the behavioral effect of insulin inactivation in so far as the previous single glucose concentration experiment indicated a selective decrease in lick burst size (5). If decreased NAc insulin signaling mediates the effect of HF-HS on lick microstructure, the sex difference may reflect a relative resistance of females to central effects of the HF-HS diet and/or a specific difference in CNS consequences of peripheral hyperinsulinemia as suggested previously by opposite correlations between insulin level and behavior (8). Going forward, additional light may be shed by determining the effect of this HF-HS regimen on NAc insulin receptor sensitivity in females, as well as testing whether the regimen occludes the effect of insulin inactivation in males as compared to females.

Sex differences in the behavioral and metabolic responses to HF-HS feeding have received relatively little experimental attention until recently. Results of the present and previous study do however align with findings obtained in other studies. For example, female rats show greater avidity for sucrose solution relative to males whether passively consumed or obtained via instrumental responding (30). Females are also relatively resistant to metabolic morbidities induced by high energy diet feeding (31–34) and fail to show diet-induced CNS effects observed in males including NAc AMPA receptor plasticity (35) and hypothalamic microgliosis (36),

Although diet-induced hyperinsulinemia and insulin receptor insensitivity have been associated with behavioral changes in lick microstructure and flavor-nutrient learning, causal connections remain to be determined. The behavioral effects do align with the finding that insulin increases striatal DA release, which is inhibited in male rats by the current HF-HS liquid supplementation regimen (2). However, a full delineation of NAc insulin signaling and related diet-dependent effects on behavior will have to disentangle the contributions of different NAc insulin receptor populations and the mechanistic and behavioral functions they regulate. Insulin receptors have been localized to DA axons, postsynaptic D1- and D2-type medium spiny neurons, and most abundantly, to cholinergic interneurons (2,3). Moreover, insulin receptor populations in NAc shell and core exert different regulatory effects on behavior, with the former having been implicated in reinforcement of flavor-nutrient learning and promoting sugar, but not non-nutritive sweetener, consumption while the latter have been implicated in diminishing passive intake of chow as well as effortful responding for food reward (2,4–6).

What is clear from the present results is that inactivation of insulin in NAc shell inhibits motivating and hedonic effects of glucose in both male and female rats, but a diet known to induce hyperinsulinemia in both sexes mimics the effect of NAc insulin inactivation in males only. Male mice have shown a similar decrease in sucrose lick microstructure parameters when maintained on a long-term HF diet (37). However, in that study the decreases were attributed to taste contrast between the home cage HF diet and sucrose, based on the observed recovery of lick microstructure parameters following a ten-day switch to control chow. That explanation seems unlikely to fully account for present results in male rats for two reasons. First, it would be challenging to explain why females are exempt from a contrast effect. Second, it was shown using the present male rat model that blockade of the receptor for advanced glycation end products, which otherwise activates inflammatory gene expression and promotes insulin resistance, reversed the effect of HF-HS on lick burst size without altering home cage HF-HS intake (8). Among the questions to be addressed going forward are the comparative time courses of HF-HS diet effects on NAc insulin receptor sensitivity and behavior in males and females, the role and possible sex differences in diet-induced signaling by adipocytes, whether insulin receptors on different cellular elements in NAc are differentially subject to diet-induced adaptations and whether those adaptations develop with the same time course.

The decreased behavioral responses to glucose in InsAb-treated rats suggests that subjects expressing naturally occurring NAc insulin resistance might seek higher glucose concentrations to attain the reinforcing effects of the commonly preferred 10%−15% concentration in non-resistant subjects (38,39). Given that only males showed an InsAb-like response to HF-HS diet it was expected that these rats would show an increase in preference ratio for 40% relative to 10% glucose. This was not seen. The decreased intake of both concentrations could reflect an insurmountable diet-induced anhedonia, common in human type 2 diabetics (40,41). The increased glucose intake and preference for 40% relative to 10% in females is in line with the observed tendency of this HF-HS diet to increase motivation for glucose (number of lick bursts) and suggests a diet-induced augmentation of the female avidity for sugar. Whether other or longer duration high energy diets lead to a more male-like profile of lick microstructure and concentration preference remains to be seen. Finally, it should be noted that all subjects in these experiments had ad libitum access to their assigned diets, were tested during the first half of the light cycle, and were presumably satiated at the time of testing. Repetition of these experiments in fasted subjects, to further prioritize nutritive value, could add important additional insight.

In the striatal slice, the DA releasing effect of insulin, examined only in male rats to date, is dependent on excitation of ChIs and α4β2 receptors (2). If the isolated and tightly controlled slice model translates to the intact brain and behavior, two predictions follow that were tested in this study. First, as with insulin inactivation, DHβE should decrease lick parameters for glucose but not saccharin. Second, DHβE should decrease acquisition of glucose conditioned flavor preference.

In female rats these predictions were borne out. DHβE strongly decreased both lick microstructure parameters for glucose and markedly decreased flavor-nutrient learning. The one caveat is that DHβE unexpectedly decreased number of lick bursts emitted for saccharin. In male rats, DHβE selectively decreased the number of lick bursts emitted for glucose and decreased flavor-nutrient learning. An effect not predicted was the enhancing effect of low dose DHβE on lick parameters for saccharin. A recent study in mice (42) revealed that ablation of D1 DA receptor-, but not D2 DA receptor-, expressing medium spiny neurons (MSNs) in NAc abolished the ability to associate a tastant with its nutritive value. An earlier pharmacological study in rats produced the compatible finding that NAc microinjection of a D1 DA receptor antagonist prevented acquisition of flavor-nutrient learning (18). Systemic pharmacology has repeatedly shown that D2 DA receptor antagonism selectively decreases lick burst size (15,16,43), while two studies have shown that D1 DA receptor antagonism selectively decreases lick burst number (14,16). Consequently, an inhibitory effect of DHβE on number of lick bursts emitted should be associated with an inhibitory effect on flavor-nutrient learning, assuming involvement of overlapping populations of D1 receptor-expressing MSNs. This was the case for both sexes.

Oral infusion of saccharin releases DA in NAc as does sugar (44,45), and present results suggest some nicotinic regulation of palatability-driven (i.e., saccharin) licking that is sex-dependent and independent of ChI insulin signaling. There are several known complexities to NAc nicotinic cholinergic function that may contribute to effects on saccharin licking and the sex difference. First, contrary to a once held belief, ChIs do not provide the only cholinergic input to NAc (46). There is also cholinergic innervation of NAc from the pedunculopontine nucleus where cholinergic neurons encode information about reward acquisition and changing reward contingencies (47), and the laterodorsal tegmental nucleus where cholinergic neurons express ghrelin receptors that contribute to nicotinic facilitation of consummatory behavior (48). Further, there are recent findings of opposing effects of ChI activation on motivation and reward that may depend on the activity state of meso-accumbens DA neurons (49). It is known that β2-containing nAChR activity facilitates low probability DA release but decreases release during DA neuron burst firing (50,51); the enhanced ability of females to release DA in response to high impulse flow has been suggested to result from a sex difference in ChI-based nicotinic regulation (52). Speculatively, a sex difference in function among ChIs that regulate the DA response to sweet taste in the absence of insulin could play a role in the opposite behavioral effects of DHβE in males and females licking for saccharin.

Among the factors that may contribute to sex differences in diet and DHβE effects, and the failure of DHβE effects to precisely follow effects of InsAb, would be sex-based differences in DA dynamics (52), differences in hormonal modulation and adipocyte signaling (e.g., 34,53), and differential effects of ChI activation in the absence of concurrent insulin signaling (54). Ultimately, a better understanding of the in vivo relationship between diet, NAc insulin signaling and nicotinic signaling as they relate to motivating and rewarding effects of nutritive and non-nutritive sweeteners may require additional investigation of diet parameters, modulatory effects of sex hormones and adipokines, and use of mouse models that lack insulin receptors in ChIs or α4β2 receptors on DA axons.

Highlights.

• Glucose concentration-response testing, using a lick microstructure assay, showed a decrease in parameters indicative of motivation (number of lick bursts emitted) and hedonic impact (average number of licks per burst) in male and female rats when insulin was inactivated in nucleus accumbens shell.

• Six weeks continuous access to a high fat-high sugar liquid (chocolate Ensure), previously shown to induce hyperinsulinemia and nucleus accumbens insulin receptor insensitivity in males, decreased lick microstructure parameters for glucose in males but not females.

• Previous findings in the striatal slice demonstrated that dopamine release in response to insulin is dependent on excitation of cholinergic interneurons and α4β2 receptors. Here, microinjection of the α4β2 antagonist, DHβE, decreased the number of lick bursts for glucose in males and number of lick bursts and lick burst size in females.

• Nucleus accumbens insulin inactivation was previously shown to decrease flavor-nutrient learning in male rats; females were not tested. In the present study, microinjection of DHβE was shown to decrease flavor-nutrient learning in both sexes.