Here, we test the relationship between D2 and/or D3 receptors and their involvement in modulating the activity of the mPFC during the processing of certain sensory stimuli. First, we tested a mouse’s ability to discriminate between a single perceptual dimension of a stimulus (i.e., smell) and for the ability to shift attention when confronted with a different perceptual dimension (i.e., touch). In each mouse, we used immunocytochemical methods to detect the expression of the immediate early gene: c-fos, a gene with low baseline expression but robust mRNA expression in neurons when activated by a large variety of stimuli, including those in the attention-set shifting tests. In each case, the number of trials, the amount of time needed for the mouse to reach a designated attention level, and the activation patterns of the AC was recorded and used to statistically evaluate the hypothesis that in reference to attention, the AC either controls “error detection,” the delay in the rate of “decision making,” or both.

Fifty adult male and female mice with postnatal ages of 60 to 90 days (13 WT, 7 D2-/-, 13 D3-/-, 17 D2/D3 double mutants) were housed in pairs in 34.0 x 16.5 x 14.5 cm plastic cages. Testing was conducted during the light phase of a 12-hour light/dark cycle (testing began at 11:00 AM). Starting seven days prior to testing, the amount of appropriated food was gradually reduced in order to lower the overall body weight to 80-85% so the mice would be motivated to forage during experimentation. Water was always provided freely in the home cage. All experiments involving the animals were approved by the Institutional Animal Care and Use Committee at Columbia University.

Apparatus

This experiment used small terracotta flowerpots (5.5 cm in diameter and 1 cm deep) as the target area where the mouse would search for food. Food pellets were baited under a certain compound-covering, such as ribbon, yarn, woodchip medium, shredded rubber gloves, beads, and plastic tops. The reward was a food pellet that the mouse was accustomed to eating in identical terracotta pots in its home cage. The rim of each pot could be scented with commercial liquid perfumed oils. The odor was refreshed before each testing session, and the choice of odors was randomized daily, to prevent the occurrence of a specific pattern.

The testing apparatus was a Plexiglas box measuring 31 x 37 x 15 cm with a transparent movable barrier (along the length of the box), separating one-third of the box from the rest, thereby creating a holding chamber (Figure 1). In each trial, two digging pots were randomly placed adjacent to each other in the main section of the box while the mouse remained in the holding chamber. The mouse was given access to the main section by raising the barrier, which was quickly replaced once the trial began.

Figure 1 . Top-down view of the attention-set shift experimental cage, made of Plexiglas. The mouse is initially held in the holding area. When a trial starts, the movable barricade is raised and the mouse gains access to two small terracotta pots, both of which are scented and covered: only one is baited.

Habituation

Prior to testing, the mouse was placed in the testing box and allowed to survey the holding chamber for 1 minute. At this time, the barrier was raised and the mouse was given access to two pots, each baited with a sliced food pellet and filled with wood chip bedding (identical to that found in its home cage). The mouse was allowed to search for food (Figure 2a) until a reliable digging behavior was acquired, which took on average, 20 minutes.

Figure 2a.. A mouse digging for food in a terracotta pot. b. After successfully obtaining the reward, the mouse is allowed to eat for 10 seconds before being removed to the holding chamber.

Testing Paradigm

Each mouse was exposed to five testing stages, which always followed the sequence: simple discrimination (SD), compound discrimination (CD), Intra-dimensional shift discrimination (IDS), IDS reversal discrimination (IDSrev), and Extra-dimensional shift discrimination (EDS). Table 1 shows a mock setup of a complete test to which a mouse could be exposed.

Table 1. A Possible Series of Stimulus Pairs for a Single Mouse *Shifting from odor to digging medium as the relevant dimension occurs in the IDSrev stage. All mice were given different, randomized test sequences to prevent the development of a behavioral pattern.

SD is a test where only one of the two distinctly scented pots is baited, but both contain the same filling medium. Initially in the SD, there are four “discovery” trials where the mouse is allowed to dig in both pots before it is removed to the holding chamber. After which, every dig in the incorrect, un-baited pot, is considered an error, and the mouse is removed to the holding chamber. On consequent trials, and after obtaining the reward, the mouse is allowed to eat for 10 seconds (Figure 2b, Video 1). The number of trials and the amount of time needed for the mouse to obtain six correct trials successively is recorded. This process is followed in the subsequent tests. Discovery trials, however, were not used as the environment, and stimuli are no longer considered to be overwhelming.

Video 1 . (Clicki mage to view or right click it to save video)

In CD, a secondary sensory dimension (i.e., touch, if scent was initially used in SD, or vice versa) was added to the paradigm. If scent was the primary stimulus, then the covering media was randomized between pots. If touch was the primary stimulus, scent was randomized between pots. Thus, the mouse is forced to rely on the relevant primary stimulus and disregard the irrelevant secondary stimulus to obtain the reward.

In IDS, the mouse is presented with a new set of odor and media stimuli while maintaining the same relevant dimension used in the CD. In the IDSrev, the relevant dimension stays the same as in the IDS. However, within the relevant dimension, the positive and negative stimuli are reversed, while maintaining identical, randomized, secondary stimuli. In the EDS, the mouse is presented with a new set of odor and media stimuli, with reversed relevant and irrelevant dimensions with respect to the IDS. Thus, in each stage, added characteristics provide an added challenge that forces the animal to pay closer attention to the sensory stimuli. Within each genotype, half the mice were shifted from odor to medium and the other half was shifted from medium to odor. The same mouse was used across test treatments from SD to EDS, after which the animal was sacrificed approximately 1 hour after the testing was completed. The order of stimuli within a dimension was randomized across subjects.

Data Analysis

During each of the discrimination tests, the total amount of time (in minutes) taken to complete six correct consecutive trials was recorded. The number of correct trials was multiplied by 10, to account for the total number of seconds each mouse spent eating during the test. Eating time was removed to give a “corrected” time, which was then divided by the total number of trials to give a response time. The means and standard deviations for response times and the total number of trials to criterion were then used as statistical data to compare the experimental results across the genotypes. Next, a one-way analysis of variance (ANOVA) with a threshold of significance of 0.05 was completed followed by a Tukey-Kramer Multiple Comparisons Test using the analysis software Instat (version 2).

Histology

An hour after the completion of testing, each mouse was given a dosage of ketamine (30mg/kg) and xylazine (7mg/kg) in a saline solution. The mouse was then perfused transcardially by inserting a solution of with a 4% formaldehyde solution directly into the heart. The brain was removed, preserved in 4% formaldehyde solution in phosphate buffer, and stored overnight. The brain was then sectioned coronally on a free-sliding microtome to a thickness of 40 ?m. For the examination of the c-fos protein immunoreactivity with fluorescence microscopy, the sections were placed in an anti-c-fos primary antibody, after which the sections were submerged in a flourescein isothiocynate-conjugated goat anti-rabbit IgG secondary antibody for 1 hour at room temperature. Sections were mounted onto DPX-coated slides, air-dried, cover slipped, and viewed microscopically using a microscope equipped with epifluorescence at 10x and 20x magnification. These observations were used to determine and to compare the extent of AC activation of the immediate early gene, c-fos, across the different genotypes.

Wild type performed with higher efficiency on the simpler tasks of SD, CD, and IDS (mean number of trials ˜ 9) but then showed a 15% gradual increase in error rates in the more attention-requiring IDSrev and EDS stages (mean number of trials ˜ 10.5), a trend which does not reach statistical significance (Figure 3).

In comparison with WT, D3 mutants exhibited a similar performance in SD and CD but performed significantly better in the more complicated IDSrev task (p < 0.03), and consistently, but non-significantly, outperformed WT in IDS and EDS tasks (Figure 3).

Figure 3 . Number of trials to criterion per genotype (mean ± SEM.) In comparison to control mice, D3 mutants performed better in the more complex shifts to IDS and EDS while making the fewest errors.

Response Time

All genotypes had more trouble adjusting to IDS and EDS shifts, as shown by the higher response times in these two particular tests in comparison with the other discriminations (Figure 4). The WT showed a gradual slowing in response time from SD to CD to IDS, an increase in speed on the IDSrev, and then a slowing with the EDS stage. Contrasted to WT, the same general trend in the overall response time is detected for in D3 mutants. The most striking difference, however, was the much slower response times of these mutants in each discrimination (Figure 4). For instance, D3 mutants performed 55% slower than WT in SD and 66% slower in the EDS discrimination. These differences were marginally insignificant for response time in the SD (p < 0.08) and were statistically significant for the CD (p < 0.03), and EDS (p < 0.03) trials. The response times of the D2 and DM did not differ from WT. Preliminary experiments on D2 mutants suggest impairments in the initial discrimination of SD. This delay is reasonable, considering the novelty of the environment and task being presented to the animal, thereby making SD the only stage where these mice tend to take extended periods in decision-making.

Figure 4. Response times for each genotype (mean ± SEM.). Overall, D3 mutants have the slowest response times.

Fos Activation

The expression of Fos immunoreactivity is correlated with neuronal activation during the task, as basal Fos levels are low (Figure 5C). There is strong activation of the AC of the post-EDS WT mouse in comparison to its basal counterpart (Figure 5C and D). This discrepancy shows that the AC is involved in the cognitive testing associated to our attention set-shifting experiment. However, the WT activation of the AC is only moderate in comparison with the strong activation of neurons in AC of D3 mutants (Figure 5E). The neuronal activation of the DM appears similar to WT. Data collected on the neuronal activation in D2 mutants is similar to the activation found in WT and is considerably weaker than that of the D3 mutants (Schmauss, unpublished data).

Figure 5a. Schematic illustration of the regional extent of the mouse medial prefrontal cortex (mPFC) which includes the anterior cingulate, the prelimbic area, and the infralimbic areas. b. Nissl stain of the mPFC which illustrates the neuronal cytoarchitechure of the AC. c. Basal Fos expression in WT where, besides a single neuron activated, there is a lack of Fos expression. d-f. The neuronal activation of c-fos in WT, D3, and DM mice, respectively, one hour after the EDS testing was completed. The scale bar in (f) is equivalent to 500 µm.

Discussion

In comparison to WT, the DM mutants show no statistically significant differences in either response times or number of trials to criterion. The lower activation of AC neurons in the DM when compared to D3 support this observation. The lack of both D2 and D3 receptors simultaneously creates a behavior that is less reliant on the AC than the D3 single mutant. The lower activation of AC neurons appears to be associated with higher error rates in harder discrimination tests, resembling the WT behavior. Reference to raw data of the DM genotype shows that these mice carried the most variability having instances where a highly accurate yet delayed D3-like behavior was present or conversely, where a quicker, more error-prone D2-like behavior was present. For instance, in IDS one DM mouse took 13 minutes to complete six trials, whereas another needed 22 minutes to complete 24 trials for criterion.

Preliminary data obtained from D2 mutants yielded results similar to those obtained with DM. Although not statistically significant, both D2 and DM mutants had trouble associated with the CD stage of the attention set-shifting experiment, making 28% and 41% more errors for DM and D2, respectively. This stage represents the first time the mouse encounters four separate sensory stimuli. Additionally, the AC activation of the D2 mutants mimics that of the WT (Figure 6Fi). It is possible that the D2 receptor is responsible for the errors associated with increasing complexities of stimulus due to the increase in difficulty that occurs with each stage progression. This can be seen in the high number of trials needed for criterion in this task in both the D2 single mutant and DM mutants, and the ease with which the D3 mutants handled this discrimination. Further studies on the D2 single mutants are necessary to clarify this issue. In addition, examination of the anatomy of the mPFC in all genotypes after the CD stage as opposed to the EDS stage may provide answers as to which neuronal systems cause the marked decrease in performance in such a simple one-dimensional task.

Neither the DM mutants nor the D2 mutants significantly differed from WT in the number of trials to criterion (Figure 3). Experimentation of the D2 genotype was carried out last and consisted of a fewer number of mice tested (seven) compared with the other genotypes. In the future, more D2 mutants must be tested to obtain an accurate estimate of their performance. Clearly, these data require further quantification, and analysis.

Comparing the neuronal activation pattern of single mutants to the activation found in the double mutant is another step towards elucidating whether both receptors have perhaps opposing actions and/or activate opposing neuronal circuits. Quantifying the number of neurons activated in each of the genotypes will allow for this precise comparison. Only at that stage could the AC be definitively assigned a cognitive role in relation to creating a higher attention rate.

Figure 6. Neuronal activation of the anterior cingulate in WT, D3, and D2 genotypes. The D2 and WT activation is considerably weaker than the activity seen in the AC of the D3 mutants (Schmauss unpublished data).