**3. Results**

In the non-choice acceptance tests, sated adult mice and rats showed very low levels of consumption of a 'bland' cornstarch emulsion, whereas intakes of the GM (mice, *F*(3,30) = 62.8, *p* < 0.001; rats, *F*(3,32) = 25.5, *p* < 0.001) and CM (mice, *p* < 0.001; rats, *p* < 0.001), as well as of the sucrose

solution (mice, *p* < 0.001; rats, *p* < 0.001), were several times higher than of cornstarch. Energy-deprived animals had a higher baseline intake of cornstarch, but consumed significantly more sucrose (mice, *F*(3,32) = 9.77, *p* ≤ 0.001; rats, *F*(3,26) = 5.5, *p* = 0.039), GM (mice, *p* < 0.001; rats, *p* = 0.0023), and CM (mice, *p* = 0.034; rats, *p* = 0.0083; Figure 1A–D). Similarly, both deprived and sated adult individuals ate more GM- and CM-enriched pellets than standard chow (sated mice: *F*(2,19) = 5.9, GM, *p* = 0.029 and CM, *p* = 0.011; sated rats: *F*(2,19) = 20.5, GM, *p* < 0.001 and CM, *p* = 0.0011; deprived mice: *F*(2,19) = 6.5, GM, *p* = 0.0058 and CM, *p* = 0.034; deprived rats: *F*(2,22) = 10.8, GM, *p* < 0.001 and CM, *p* = 0.0442; Figure 1E–H). Adolescent and aged sated mice and rats (Figure 2A,B,E,F) given episodic 2-h access to one of the solutions, consumed more GM (adolescent mice, *F*(3,35) = 42.7, *p* < 0.001; rats, *F*(3,36) = 16.9, *p* < 0.001; aged mice, *F*(3,29) = 31.2, *p* < 0.001; rats, *F*(3,29) = 18.9, *p* < 0.001), CM (adolescent mice, *p* < 0.001; rats, *p* < 0.001; aged mice, *p* < 0.001; rats, *p* < 0.001) and sucrose (adolescent mice, *p* < 0.001; rats, *p* < 0.001; aged mice, *p* < 0.001; rats, *p* < 0.001) than cornstarch.

When given a 2-h episodic choice between GM and CM, all age cohorts of rats (adolescent, *p* < 0.001; adult, *p* < 0.001; aged, *p* < 0.001) and adult and aged mice (*p* = 0.012 and 0.011, respectively) preferred GM (Figure 2C,D,G,H and Figure 3A,B). During a brief, 30-min exposure to both GM and CM in cages equipped with lickometers, adult rats exhibited a more robust response to GM cumulatively over that period (*p* = 0.01) as well as during the first (*p* = 0.037) and second (*p* = 0.05) 5-min time interval of the meal (Figure 3C,D). There was a trend approaching significance (*p* = 0.088) toward an increase in the cluster number (number of licking bouts) of GM over CM, and a significantly greater cluster length of each GM than CM bout (*p* = 0.022; Figure 3E,F). In choice experiments involving GM- and CM-enriched chow, adult and aged rats (*p* = 0.009 and 0.023, respectively) and adult mice (*p* = 0.028) preferred GM chow, whereas in aged mice, a trend toward GM preference was detected (*p* = 0.059) (Figure 4A,B). Adult rats given a 72-h uninterrupted access to a choice between GM and CM chow preferred GM chow (*p* < 0.001), while both GM (*p* = 0.015) and CM pellets (*p* < 0.001) were preferred over standard food during a similar time of exposure (Figure 4C).

Real-time PCR analysis after consumption of the two milk formulations (GM: 19.27 +/− 0.18 g; CM: 18.44 +/− 0.17 g) revealed that GM upregulated in the nucleus accumbens PNOC (*p* = 0.0164), ORL1 (*p* = 0.0042), Cx36 (*p* = 0.0017), GLP1R (*p* = 0.0015), MC4R (*p* = 0.002), OXT (*p* < 0.001), and GHSR (*p* < 0.001) genes, whereas mRNA levels of PENK were lower (though it did not reach significance with a *p* value of 0.01), compared with CM consumption. In the hypothalamus, MOR (*p* = 0.045) and KOR (*p* = 0.017) transcript levels were higher after GM consumption, and in the brain stem there was a trend toward upregulation of the MC4R (*p* = 0.099) and the MC3R was upregulated (*p* = 0.0275; Figure 5). Compared to water controls, in the nucleus accumbens, GM affected expression of ORL1 (*p* = 0.012), Cx36 (*p* = 0.0052), GLP1R (*p* = 0.0042), MC4R (*p* = 0.0053), OXT (*p* = 0.0149), and GHSR (*p* < 0.001); in the hypothalamus, ORX (*p* = 0.0164), KOR (*p* = 0.0399), and MC4R (*p* = 0.0403). On the other hand, hypothalamic expression of the MC4R gene was elevated by CM intake (*p* = 0.041; Figure 5).

**Figure 1.** Episodic 2-h consumption of individually presented (acceptance) cornstarch, sucrose, GM, and CM isocaloric solutions (**A**–**D**), and of standard, GM- and CM-enriched chow (**E**–**H**) in sated (nondeprived) and energy-deprived mice (left panel) and rats (right panel).\*, *p* ≤ 0.05; \*\*, *p* ≤ 0.01; \*\*\*, *p* ≤ 0.001.

**Figure 2.** Episodic 2-h consumption of individually presented cornstarch, sucrose, GM, and CM isocaloric solutions (**A**,**B**,**E**,**F**: acceptance) and simultaneously given GM and CM (**C**,**D**,**G**,**H**: preference) in adolescent and aged sated mice (left panel) and rats (right panel). \*, *p* ≤ 0.05; \*\*\*, *p* ≤ 0.001.

**Figure 3.** Episodic consumption of simultaneously presented GM and CM over 2-h in sated mice (**A**) and rats (**B**), lickometer activity during a 30-min exposure ((**C**): 0–30 min; (**D**): 5-min intervals), the number of GM over CM licking bouts (cluster number) (**E**), and the cluster length(s) of each GM and CM bout (**F**) in sated rats. \*, *p* ≤ 0.05; \*\*\*, *p* ≤ 0.001.

**Figure 4.** Consumption of simultaneously presented GM- and CM-enriched chow in adult and aged sated mice (**A**) and rats (**B**) over 2 h and simultaneously presented pellets (standard vs. GM; standard vs. CM, and GM vs. CM) over 72 h in adult rats (**C**).\*, *p* ≤ 0.05; \*\*, *p* ≤ 0.01; \*\*\*, *p* ≤ 0.001.

**Figure 5.** Relative expression of feeding-related genes in the nucleus accumbens (**A**), hypothalamus (**B**), and brain stem (**C**) of mice maintained for 24 h on GM or CM. Water served as a baseline tastant. a—significantly different from the water group; b—significantly different from the CM group. Analysis performed with ANOVA followed by Bonferroni's test and corrected for multiple comparisons.

#### **4. Discussion**

Enhanced motivation to eat in the absence of an immediate need to replenish calories or continuation of a meal beyond levels that restore energy balance typically occur when an individual is given access to food that is highly palatable. In laboratory animal models, similarly to what is observed in humans, a variety of tastants are perceived as palatable. Those include ingestants whose palatability is derived mainly from the flavor and/or postabsorptive effects of either a single macronutrient (e.g., sucrose-sweetened solutions) or from the complex contribution of multiple nutritive components (e.g., in meat rich in protein and fat) [30–32]. Calorie density of food (especially when coupled with high energy needs of an organism) is an additional factor that affects the liking of and preference for a given food [15,33].

The current set of studies show that both GM and CM and milk-enriched solid diets are highly palatable. In no-choice acceptance paradigms, energy non-deprived rats and mice of all age groups (adolescent, adult, and aged) consumed GM and CM as avidly as the calorie-matched 15% sucrose solution (used here as a positive control for a highly palatable tastant in rodents (for review, see [30]), while ingesting only minimal amounts of the 'bland' cornstarch. A similar phenomenon was observed in energy-deprived animals, although the amount by which GM, CM, and sucrose intakes exceeded that of cornstarch was not as pronounced as in sated rodents. That was due to the vigorous energy deficit-driven consumption of cornstarch and a 'ceiling effect' that prevents ingestion of large amounts of the solutions during the brief refeeding period. Importantly, GM and CM enrichment of laboratory chow stimulated intake in both hungry and sated animals well above the level of standard pellets. It indicates that both GM- and CM-derived palatability is a generalized phenomenon, not limited to liquid milks, but extending to solid foods that contain milk powder. This is in concert with the ability of other palatable tastants (including, but not limited to, fat, sucrose, and select amino acids) to have a positive gustatory effect when presented as a component of both liquid and solid foods [34]. The fact that not only adolescent and adult animals, but also the aged ones, readily consume GM and CM suggests that age-related decline in hedonic processing [22,35–37] does not completely abolish a drive to eat milk-based diets. Instead, a slightly depressed intake of GM and CM at an old age parallels that reported for sweet solutions, as shown here and by other authors [38–40]. This finding is particularly relevant from the standpoint of being able to use palatable GM or CM as nutritionally superior alternatives to, e.g., sucrose-sweetened tastants in aged individuals [23]. That adolescent rodents also consume large quantities of both milk types indicates that prolonged dietary habituation is not required to develop the liking of either GM or CM. In fact, the amounts of GM and CM ingested by juveniles were as high as the volume of sucrose (readily consumed in large quantities by young animals, e.g., see [41]) even though the individuals had had only two brief exposures to these solutions prior to the experiment.

The single-tastant scenarios above strongly suggest a high acceptance level for both GM and CM indicating they are palatable, but as these no-choice paradigms produced fairly similar feeding responses, choice studies were needed to define relative preference for these two milk types. Simultaneous 2-h exposure to two bottles containing GM and CM showed that adult and aged mice and rats as well as adolescent rats exhibit a marked preference for GM (adolescent mice were the only cohort in which GM and CM were iso-palatable). The preference for GM did not appear to be related to whether the animals' pre-exposure to the specific diets was simultaneous (such as in adolescents and adults) or sequential (aged rodents). This finding was further expanded by employing the 30-min lickometer analysis in adult rats. It showed approximately four times as many licks at the bottle containing GM compared to CM during the first 5 min of the meal, and twice as many licks at the GM bottle in the subsequent 5-min interval. Overall, the licking activity at both bottles occurred within the same timeframe with neither milk type being ingested in a prolonged fashion. It increases our confidence in that motivation to consume palatable GM rather than maintenance of a meal (due to, e.g., delayed satiation [42]) is the main reason for avid intake of GM. The analysis of the licking bouts provides additional support for this notion. The cluster number (total number

of bouts) neared significance for GM, possibly reflecting the incentive motivational properties of the food stimulus; importantly, the relationship of motivation and this measure reflects post-ingestive negative feedback [43–47]. On the other hand, the average cluster length—significantly greater for the GM formulation—typically parallels the hedonic properties (mainly, orosensory pleasure) of ingestive stimuli (as reviewed, e.g., in [45]). In this case, it is the length of clusters that appears to be the main driver for the preference for GM. A good example of the significance of licking bout length versus number in the context of neural regulation of food intake comes from studies on the endogenous opioid system. Ostlund et al. found that mu opioid receptor (MOR) knockout (KO) mice show alterations in sucrose licking: while energy-deprived wild-type mice increased burst length, relative to the nondeprived condition, this aspect of licking was insensitive to changes in food deprivation in MOR KOs. The rate of sucrose and sucralose licking in KOs was lower than in wildtype animals, providing evidence that the MOR was involved in processing palatability [48]. Mendez and colleagues reported that proenkephalin (PENK) KOs given a sucrose solution exhibited fewer bouts of licking (though the length did not differ) than wild type controls, indicating a diminished motivation to eat [46]. Finally, studies on the involvement of nociceptin/orphanin FQ (NOC) revealed that NOC administration initiates new bouts of licking for sweet solutions, which is in line with the notion of its potential relationship to motivational aspects of feeding. Interestingly, energy-deprived NOP KO mice given sucrose showed longer bouts of licking than wild types, suggesting that, under hungry conditions, NOC may also affect hedonics of consumption [49].

The notion that satiety is not delayed by GM intake is supported by the experimental work exploring satiating effects of a CM- versus GM-based meal in humans. In their study, Rubio-Martín et al. presented healthy adults with GM-based or CM-based breakfast after an overnight fast and obtained blood samples and appetite ratings from the subjects just before and up to 5 h after completion of the meal. They found that that the 'desire to eat' rating was significantly lower and hunger rating tended to be lower after the GM breakfast. Interestingly, the area under the curve (AUC) for a satiety hormone GLP-1 was inversely associated with the AUChunger and AUCdesire-to-eat after the GM meal [50].

The aforementioned data obtained in human observations combined with the current results of our experiments in animal models suggest that even though composition differences between GM and CM are relatively minor, they are sufficient to significantly affect appetite-related parameters. It remains to be elucidated whether these effects are produced by a specific macronutrient component, a combination of nutritive components, and/or some physico-chemical characteristics of each milk type (e.g., micelle structures in GM vs. CM differ in diameter, hydration, and mineralization) [51].

The analysis of mRNA levels of feeding-related genes sheds more light on neural processing underlying enhanced preference for GM over CM. One of the most striking outcomes is the fact that, unlike in the NAcc, which showed an increase in multiple mRNA profiles after GM over CM, there are relatively few significant differences in gene expression in the hypothalamus and brain stem. Those two brain areas serve as the foundation for the control of energy homeostasis and consumption-related changes in the internal milieu associated with plasma osmolality, stomach distension, and defense from exposure to food-borne toxins [52]. In this network, the brain stem acts as the relay station between the periphery and the central nervous system, whereas the hypothalamus plays a endocrine role (by releasing, e.g., anorexigenic hormones, such as oxytocin (OXT) via the neurohypophysis) and innervates a number of central target sites (it includes the reciprocal connectivity with the brain stem, as well as multiple pathways with, among others, nigrostriatal and hippocampal structures). It is noteworthy that, despite the same level of intake of GM and CM over the 24-h period, the hypothalamic expression of NPY and orexin (ORX) was lower in the GM group. Both ORX and NPY in the hypothalamus enhance consumption chiefly by increasing hunger and motivating intake of energy-dense tastants [53,54]. Thus, these data suggest that enhanced preference for GM over CM does not stem from the stimulation of neural mechanisms that lead to hunger-driven feeding. In line with the aforementioned conclusion from feeding experiments that the increased preference for GM vs. CM in choice scenarios is unlikely to be related to suppressed satiety signaling, we found that the brain stem expression of satiation promoting melanocortin receptors [55,56] is elevated after consumption of GM (it remained the same in the hypothalamus). This change in the receptor mRNA level coupled with the lack of a difference in the melanocortin ligand precursor gene expression (proopiomelanocortin, POMC) as well as in the anorexigenic OXT gene [28,57] suggests the lack of impairment in central satiety processing after GM (and, surprisingly, even a somewhat greater sensitivity of the molecular network promoting satiety in response to GM consumption).

Interestingly, the hypothalamic genes whose expression was elevated by GM intake were those encoding the MOR and kappa (KOR) opioid receptors (MOR and KOR brain stem and accumbal mRNA levels were also higher, though the difference did not reach statistical significance). Furthermore, in the NAcc, we found overexpression of genes coding for opioid-related NOC and this peptide's receptor, ORL1. Opioid receptors are directly implicated in the regulation of feeding for reward [14,42]. They are part of a dispersed network that includes the NAcc as one of the key sites mediating hedonic aspects of eating behavior. They are also expressed throughout the 'homeostatic' components of the feeding-related circuit [16], including the hypothalamus and brain stem, where they are theorized to promote excessive consumption of palatable tastants by delaying meal termination. The magnitude at which opioid receptor agonists, such as butorphanol tartrate, dynorphin and beta-endorphin, stimulate consumption parallels the relative palatability of foods [14,58]. Conversely, opioid receptor antagonists, e.g., naltrexone and naloxone, are particularly effective at reducing intake of tasty ingestants [59]. Hence, higher expression of the MOR and KOR mRNA after GM is in line with the observed preference for the GM over CM. Changes in expression of additional NAcc genes that underscore the functional relationship between GM intake and reward processing include upregulation of Cx36 mRNA, as Cx36 ensures proper synchrony of dopaminergic pathways [60], and of the growth hormone secretagogue receptor (GHSR) mRNA, considering that the GHSR in the NAcc has been found to mediate hedonics of ingestive behavior [61]. Again, as in the case of the hypothalamic gene expression analysis, genes encoding molecules that promote satiety—such as OXT, melanocortin receptor 4, and glucagon-like peptide-1 receptor [62]—were upregulated after GM, which points to the heightened reward processing rather than impaired satiation as the factor propelling preference toward GM over CM.
