Behavioral and Neural Characterization of Conditioned Flavor-Taste Preferences

Many animals, including humans, choose their source of nutrition based on the nutritive value and the flavor (i.e, odor, taste and texture) of available foods. Sweet taste is one of the more potent orosensory stimuli that contributes to food choice. Laboratory animals develop preferences for neutral or aversive tastes and flavors by associating them with the taste of sugars and non-caloric sweeteners. Learning how these preferences develop would aid in understanding how and why specific foods are selected over others. Given the wide availability and variety of sweetened foods and beverages in modern society, the formation and persistence of learned food preferences by consumers may contribute to health issues such as diabetes and obesity. Conditioned flavor-taste preference (CFTP) learning is a form of associative learning in which a rat comes to prefer a neutral flavor paired with a preferred taste. Experimentally, one flavor (the conditioned stimulus or CS+; e.g., cherry or grape Kool-Aid) is paired with the sweet and highly preferred taste of fructose (F; the unconditioned stimulus or US) while a second flavor (the CS-) is paired with the less preferred taste of saccharin (S) on 1-bottle conditioning days (CS+/F or CS-/S). The acquisition of the learned preference is then assessed with a 2-bottle preference test in which both flavors mixed with saccharin (CS+/S and CS-/S) are presented simultaneously. While CFTP learning is well known, it has not been well characterized. The olfactory and gustatory associative brain regions necessary for CFTP learning are unknown. Dopamine receptors have been implicated, but otherwise it is not known which neurotransmitters or receptors mediate CFTP. In order to identify the associative neural substrates that are involved in CFTP learning, three approaches were taken; behavioral, pharmacological and molecular assays. To precisely characterize the behavioral acquisition and expression of a CFTP, lickometers were used to determine the pattern of drinking in rats. During 1 or 2-bottle preference tests, total intake, bout size, bout number, lick rate and first minute licks were analyzed. The pattern of drinking was examined under 3 conditions: 1. During expression of unconditioned preferences for 8% fructose over 0.2% saccharin. The unconditioned preference for fructose over saccharin was slow to develop, but was seen in significantly greater total intake, bout size, and first minute licks for fructose by the fourth preference test. 2. During the pairing of Kool-Aid flavors with either 8% fructose or 0.2% saccharin. CS-/S total intake and bout size was significantly greater than CS+/F during conditioning, but a preference for the CS+ flavor was seen in the third and fourth 2-bottle preference test days with significantly greater total intake and bout size of CS+/S vs. CS-/S. 3. During long-term presentations of Kool-Aid mixed with two different concentrations of saccharin (0.2% vs 0.05%) as the unconditioned stimuli. Total intake, bout size and bout number were significantly greater for the flavor mixed with the high concentration of saccharin over the low concentration of saccharin. during conditioning. During 2-bottle preference tests when both flavors were mixed with the low concentration of saccharin, total intake and bout size were significantly greater for the CS+. The increases in lick rate and bout size observed in all 3 experiments suggest that fructose is more palatable than saccharin, and a high concentration of saccharin is more palatable than a low concentration. A change in relative palatability of the Kool-Aid flavors is conditioned by association with the high palatability tastes; greater intake of the conditioned flavor is mediated by increased bout size. These results suggest that flavor preference learning interacts with both orosensory processes and satiety processes (i.e. prolonged bout size) to elevate intake of the preferred flavor. The N-methyl-D-aspartate (NMDA) glutamate receptor (NR) is a candidate mediator in olfactory and taste learning (Barkai and Saar, 2001; Jimenez and Tapia, 2004). To determine if NR is involved in CFTP, systemic MK-801, a non-competitive NR antagonist, was administered prior to conditioning and prior to expression. To determine the glycinergic contribution to NR activation in CFTP, systemic D-cycloserine, an agonist at the NR glycine-binding site, was administered prior to conditioning and reversal learning. While vehicle-treated rats acquired a preference for CS+/S over CS-/S, CFTP learning was completely blocked in MK-801-treated rats. The effect of MK-801 was specific to CFTP acquisition, because follow-up experiments demonstrated that MK-801 did not induce a conditioned taste aversion, cause state-dependent learning, or affect CFTP expression. In a second approach, rats were injected with DCS (15 mg/kg) 60 min prior to daily conditioning. In contrast to MK-801, administration of DCS prior to conditioning enhanced CFTP learning (but not reversal conditioning). These results demonstrate that NR neurotransmission is critical for CFTP learning. Furthermore, enhancement of CFTP learning by DCS suggests that endogenous levels of glycine or D-serine may be a limiting factor in CFTP learning. To determine the activation of neural populations during associative CFTP learning, c-Fos immunohistochemistry was used to illuminate the differential patterns of cellular activation. To eliminate the potential confounds of food restriction, restricted drinking sessions and potential postingestive effects that may effect c-Fos activation, rats were conditioned using a highly preferred concentration (0.2% saccharin) and a lesser preferred concentration of saccharin (0.05% saccharin) as the unconditioned stimuli. In order to standardize exposures, stimuli were applied by intraoral infusion. C-Fos immunolabeling was visualized within the brain after an intraoral infusion of either CS+ or CS- flavors (e.g. grape and cherry Kool-Aid) in combination with greater or lesser preferred taste US (e.g. 0.2% or 0.05% saccharin). Neuronal activation was assessed in forebrain sites (e.g. gustatory cortex, amygdala, and lateral hypothalamus). There was no difference in intraoral intake between all experimental groups in both conditioned and unconditioned rats, and extensive c-Fos activation was evoked in the olfactory, gustatory and learning relays of all experimental groups. Analysis of the differential patterns of c-Fos immunolabeling among unconditioned rats revealed a significant increase in c-Fos immunolabeling in the basolateral nuclei of the amygdala after intraoral infusions of dH2O compared to unconditioned rats after intraoral infusions of CS+/0.2% saccharin or CS-/ 0.055 saccharin. Therefore, the basolateral amygdala may be involved in the unconditioned response to sweetened flavors, or in the association of flavor with sweet tastes. Among conditioned rats, there was a trend towards greater in c-Fos immunolabeling in the lateral habenula of rats after intraoral infusions of CS+ vs the CS- or saccharin alone. Therefore, the lateral habenula, which is part of the accumbens-ventral tegmental area reward pathway, may be involved in the discrimination of learned preferences.