Fatty acid synthase as a factor required for exercise-induced cognitive enhancement and dentate gyrus cellular proliferation

Abstract

Voluntary running is a robust inducer of adult hippocampal neurogenesis. Given that fatty acid synthase (FASN), the key enzyme for de novo fatty acid biosynthesis, is critically involved in proliferation of embryonic and adult neural stem cells, we hypothesized that FASN could mediate both exercise-induced cell proliferation in the subgranular zone (SGZ) of the dentate gyrus (DG) and enhancement of spatial learning and memory. In 20 week-old male mice, voluntary running-induced hippocampal-specific upregulation of FASN was accompanied also by hippocampal-specific accumulation of palmitate and stearate saturated fatty acids. In experiments addressing the functional role of FASN in our experimental model, chronic intracerebroventricular (i.c.v.) microinfusions of C75, an irreversible FASN inhibitor, and significantly impaired exercise-mediated improvements in spatial learning and memory in the Barnes maze. Unlike the vehicle-injected mice, the C75 group adopted a non-spatial serial escape strategy and displayed delayed escape latencies during acquisition and memory tests. Furthermore, pharmacologic blockade of FASN function with C75 resulted in a significant reduction, compared to vehicle treated controls, of the number of proliferative cells in the DG of running mice as measured by immunoreactive to Ki-67 in the SGZ. Taken together, our data suggest that FASN plays an important role in exercise-mediated cognitive enhancement, which might be associated to its role in modulating exercise-induced stimulation of neurogenesis.

Figure 1. Experimental design.

Experiment 1. Mice were allowed to run for 28 days and were then sacrificed followed by the analysis of FA content in the brain and of the expression of genes encoding metabolic enzymes. Experiment 2. Mice were allowed to run for 33 days and given chronic i.c.v. injections of C75 (▾) at the indicated times. On day 29^th of running, Barnes maze behavioral testing was initiated, which consisted of five days. The complete running period spanned the entire experimental design for 33 days.

Figure 2. Voluntary running increases fasn gene expression and the content of PA and SA in the hippocampus, but not in cortex, or cerebellum.

(A) Increase in fasn gene expression by running. Amplification plots were produced to calculate the threshold cycle (Ct) followed by evaluation of the changes in the expression of target fasn relative to our reference gene, gapdh, using our java desktop application DeltaDeltaCT_v02 (see Materials and Methods). Data are expressed as the mean value using the estimated 95% Ct, represented as the mean ± SEM. Bonferroni post-testing identified a specific significant increase in hippocampal fasn mRNA levels in running versus sedentary mice (***P<0.001). Sed - sedentary, Run - running. (B) Running-dependent elevation of PA and SA in the hippocampus. Quantitation of PA and SA was carried out using each peak’s height area and the sum of all peaks from the GC/MS data. Results were expressed as a percent of total of all peaks (mean ± SEM), N = 6. Data were analyzed with Two-Way RM ANOVA followed by Bonferroni post-testing; **P<0.01 indicates specific statistically significant differences, as compared with corresponding Sedentary controls. (C, D) The expression of elovl1 and elovl6 in the brain is not potentiated by running. Amplification plots were produced to calculate the threshold cycle (Ct) followed by evaluation of the changes in the expression of target elovl 1 and elovl 6, respectively, relative to our reference gene, gapdh, using our java desktop application DeltaDeltaCT_v02 (see Materials and Methods). Data are expressed as the mean value using the estimated 95% Ct, represented as the mean ± SEM.

Figure 3. Identification of FA that influence the separation between Running and Sedentary groups in the neurolipidomics analysis.

Ellipses of the score plots (Left) generated after PCA analysis of variations in FA data sets (N = 6 per brain region) including corresponding loading plots (Right), in the hippocampus, cerebellum, and cortex of sedentary (blue circle) or running (red circle) mice. The ellipses of the score plots illustrate 95% confidence region of the groups.

Figure 4. Effects of C75 on searching strategy during Barnes maze acquisition and LTM testing.

Acquisition and LTM data depicting the particular searching strategies, random (black circle), serial (empty circle), or spatial (red circle), used to search for the escape hole within the Barnes maze as observed for treated and control mice of both Sedentary and Running groups VHL-S (A), VHL-R (B), SH-S (C), SH-R (D), C75-S (E) and C75-R (F). Graphed data represent the % of mice per session (consisting of 4 trials per day) in each group using selected searching strategies throughout acquisition and LTM testing. Two way RM ANOVA and Bonferroni post-testing (*P<0.05, **P<0.01, ***P<0.001) indicates specific statistically significant differences between groups in terms of the searching strategies preferred for searching (*-spatial vs. random; ^#-spatial vs. serial, ^&- serial vs. random).

Figure 5. FASN inhibition in the forebrain affects escape latency and spatial discrimination in the Barnes maze.

Graphed data depicting results obtained from acquisition training and LTM tests in the Barnes maze were given to Sedentary (punctuate line) and Running (solid line) mice injected with VHL (A), SH (B) or C75 (C). Each session during acquisition and LTM testing consisted of 4 training trials for which the latency in min to find the escape hole in the Barnes maze. Specific statistically significant differences between groups identified by Bonferroni post- testing specific are depicted (*P<0.05, **P<0.01, ***P<0.001). Results are presented as means ± SEM.

Figure 6. C75 affects LTM of spatial discrimination in the Barnes maze.

Bar graphs depicting the preference of mice to poke in holes within the T, NT-L, NT-R, and NT-O quadrants of the Barnes maze. Data for each group were analyzed by One-Way ANOVA coupled to multiple post-testing analyses. Statistically significant differences between groups were observed according to the number of pokes each animal executed to find the target hole identified significant differences. Data are expressed as the mean number of hole pokes ± SEM. VHL-S, VHL-R, SH-S, SH-R, and C75-S, but not C75-R, displayed significant spatial discrimination between the T vs. NT-L (***P<0.001), NT-R (⁺⁺⁺ P<0.01), and NT-O (@@@P<0.001) during the LTM test.

Figure 7. Immunohistochemistry for Ki-67+ in the SGZ of control versus treated mice.

Injected sedentary (VHL-S, C75-S) and running (VHL-R, C75-R) groups (n = 6 per group), as well as both sham controls (SH-S and SH-R; n = 5 per group), were sacrificed at the end of the four weeks of the sedentary or running period of our experimental design and their brains were used for brain Ki-67+ immunohistochemistry. (A) Representative photomicrographs show characteristic Ki-67+ clusters of round to oval-shaped cells (arrows) in the SGZ of the DG were observed to be distributed randomly in both blades of the DG. GCL, granule cell layer; SGZ, subgranular zone; H, hilus. The scale bars represent 20 µm. (B) Comparison between the Sedentary and Running groups with different treatments: C75, VHL, and Sham. Bonferroni post-testing ***P<0.001 revealed statistically significant differences among the groups, indicating that running increased Ki-67+ labeling, an effect that was blocked by C75 treatment. Error bars indicate means ± SEM.