Article Figures & Data
Figures
- Figure 1.
EE–EGCG treatment effects on young Ts65Dn mice deficits in spatial learning, reference memory and cognitive flexibility. A, Heat map representing the accumulated swimming trajectories of mice from the different experimental groups across acquisition, removal, and reversal sessions in the MWM. Periphery and center zones are depicted at upper left. Color code is represented on the right, with red corresponding to the most visited zones and black to less visited or nonvisited zones. Learning curves are represented in latency (s) to reach the escape platform (B), Gallagher distance (accumulated distance to the goal in cm; C), and thigmotaxis (percentage of time spent on the periphery; D) during the acquisition sessions. E,F, Boxplots of the distribution of latency to the first entry to the platform area and the Gallagher distance of the four experimental groups in the removal session. Dots indicate the values of each individual mouse. Reversal learning curves are represented in latency (s; G), Gallagher distance (H), and thigmotaxis (I). A1-5, acquisition sessions 1–5; R1-3, reversal sessions 1–3 with four trials per day; REM, removal session. Data in B, C, D, F, G, and H are mean ± SEM. Data in B and F were analyzed by a censored model, which considered 60 s as the maximum trial duration. Data in C, D, G, and H were analyzed with ANOVA repeated measures, and data in E were analyzed by one-way ANOVA. In all cases, Tukey multiple post hoc comparisons corrected with Benjamini–Hochberg were used. Even if all groups were considered for multiple comparisons, the figure reports only statistically significant differences of the following relevant contrasts of interest: WT versus TS; TS versus TS EE-EGCG; WT versus TS EE-EGCG. *p < 0.05, **p < 0.01, ***p < 0.001.
- Figure 2.
Supervised PCA of the experimental groups during the acquisition sessions revealed that EE-EGCG treatment improves global learning in Ts65Dn mice. A, Trajectories connect group performance (medians) along the five learning sessions (labeled with respective numbers) in the space defined by PC1 and PC2, which consist of linear combinations of the original variables. B, Variable directionality in the PCA space. Arrows represent the direction with respect to PC1 and PC2. Variables reaching the unit circle belong to variables that are well represented by the two principal components. Contribution of variables to PC1 (C) and PC2 (D) in percent. PC1 receives a similar contribution from all classic variables used to assess learning and can thus be understood as a composite learning variable. E, Boxplots of PC1 distribution for each experimental group on sessions 1 and 5 of the acquisition phase. Box plot horizontal lines, group median; box edges, 25th and 75th percentiles; whiskers, minimum and maximum values to a maximum of 1.5 times the interquartile distance from the box. More extreme values are individually plotted. Only relevant comparisons are reported in the figure for the sake of clarity (WT versus TS; TS versus TS EE-EGCG; WT versus TS EE-EGCG), even if all groups were considered for the permutation test. *p < 0.05, **p < 0.01,***p < 0.001. The benefits of the EE-EGCG treatment on Ts65Dn learning explain the displacement of this group toward more positive values along the PC1.
- Figure 3.
Supervised PCA of the experimental groups during the reversal sessions revealed poorer cognitive flexibility of untreated Ts65Dn mice. A, Trajectories of medians along the reversal session (accordingly labeled) on the space formed by PC1 and PC2. B, Variable directions on the PCA space defined by PC1 and PC2. Variables reaching the unit circle belong to variables that are well represented by the two principal components. Bar plots represent the contribution of variables to PC1 (C) and PC2 (D) in percent. E, Box plots of PC1 distribution for each experimental group on sessions 1 and 3 of the reversal phase. Box plot horizontal lines, group median; box edges, 25th and 75th percentiles; whiskers, minimum and maximum values to a maximum of 1.5 times the interquartile distance from the box. More extreme values are individually plotted. Only relevant comparisons are reported in the figure for the sake of clarity (WT versus TS; TS versus TS EE-EGCG; WT versus TS EE-EGCG), even if all groups were considered for the permutation test. *p < 0.05, **p < 0.01,***p < 0.001. The shift of mouse groups toward positive values of PC1 represents the increased cognitive flexibility of the groups, with EE-EGCG–treated Ts65Dn mice attaining higher values than their untreated counterparts.
- Figure 4.
Effect of treatment on DI in the novel object recognition test. Left, diagram of the apparatus used for the novel object recognition test. Right, boxplots of the distribution of the DI (%) among the experimental groups. Dots, DI measure from each individual mouse; dashed lines, group means; continuous lines, group medians. Data were analyzed using one-way ANOVA and Tukey multiple post hoc comparisons corrected with Benjamini–Hochberg. Even if all groups were considered for multiple comparisons, the figure reports only statistically significant differences of the following relevant contrasts of interest: WT versus TS; TS versus TS EE-EGCG; WT versus TS EE-EGCG. *p < 0.05, **p < 0.01. Ts65Dn mice show a trend toward a reduction in DI (p = 0.08). EE-EGCG treatment improves DI score in Ts65Dn mice but worsens performance in WT mice.
- Figure 5.
Ts65Dn mice show a reduction in dendritic spine density in DG and CA1, and EE-EGCG treatment ameliorates this deficit in CA1. Left, dorsal hippocampal region of a Golgi preparation illustrating dendrites from CA1 and DG subregions; scale bar represents 10 μm. A,B, Boxplots of the distribution of dendritic spine density (spines per micrometer) in DG and CA1 among the experimental groups. Dots, repeated values from individual mice (two to three dendrites per slice, three dorsal hippocampal slices per brain, five to six mice per experimental group); dashed lines, group means; continuous lines, group medians. Data were analyzed with a linear mixed model, which included experimental group as a factor and mouse as a random effect. F test was used to test the global hypothesis. Post hoc tests were applied for the following contrasts of interest: WT versus TS; TS versus TS EE-EGCG; WT versus WT-EE-EGCG. *p < 0.05, ***p < 0.001. Ts65Dn mice show a significant reduction in spine density in both DG and CA1. EE-EGCG treatment increases dendritic spine density in Ts65Dn CA1 and decreases this parameter in WT.
- Figure 6.
EE-EGCG effects on VGLUT1/VGAT puncta in DG and CA1. Box plots of the distribution of VGLUT1/VGAT ratios of different puncta density and percentage of area among the experimental groups at DG and CA1. A, C, VGLUT1/VGAT ratio of puncta density (puncta per square micrometer). B, D, VGLUT1/VGAT ratio of percentage of area occupied. Dots, repeated values from individual mice; dashed lines, group means; continuous lines, group medians. Data were analyzed with a linear mixed model, which included experimental group as a factor and mouse as a random effect. F test was used to test the global hypothesis. Post hoc tests were applied for the following contrasts of interest: WT versus TS; TS versus TS EE-EGCG; WT versus WT-EE-EGCG. *p < 0.05, **p < 0.01, ***p < 0.001.
Tables
- Table 1.
Single-variate MWM multiple post hoc comparisons with Benjamini–Hochberg correction.
95% confidence interval Figure Variable Phase Contrast Data structure Type of test Estimated mean difference SE Lower Higher p-value Fig. 1B Latency ACQ Overall effect Continuous variable Tobit model χ2(3) = 42.24 <0.001 Fig. 1B Latency ACQ TS_WT Continuous variable Tobit model 51.262 11.28 29.146 73.377 <0.001 Fig. 1B Latency ACQ TS EE-EGCG_TS Continuous variable Tobit model –23.839 12.98 –49.280 1.602 0.028 Fig. 1B Latency ACQ TS EE-EGCG_WT Continuous variable Tobit model 27.423 11.40 5.066 49.780 0.005 Fig. 1C Gallagher distance ACQ Overall effect Continuous variable ANOVA repeated-measures F-test F = 13.636 <0.001 Fig. 1C Gallagher distance ACQ TS _WT Continuous variable ANOVA repeated-measures F-test 1903.3 397.6 1124 2682.6 <0.001 Fig. 1C Gallagher distance ACQ TS EE-EGCG_TS Continuous variable ANOVA repeated-measures F-test –896.3 427.9 –1734.9 –57.6 0.043 Not shown Gallagher index ACQ Overall effect Continuous variable ANOVA repeated-measures F-test F = 10.226 <0.001 Not shown Gallagher index ACQ TS _WT Continuous variable ANOVA repeated-measures F-test 17.217 3.981 9.41424 25.01976 <0.001 Not shown Gallagher index ACQ TS EE-EGCG _TS Continuous variable ANOVA repeated-measures F-test –8.903 4.284 –17.2996 –0.50636 0.045 Fig. 1D Thigmotaxis ACQ Overall effect Continuous variable ANOVA repeated-measures F-test F = 9.22 <0.001 Fig. 1D Thigmotaxis ACQ TS_WT Continuous variable ANOVA repeated-measures F-test 21.172 5.629 10.13916 32.20484 <0.001 Fig. 1D Thigmotaxis ACQ TS EE-EGCG _TS Continuous variable ANOVA repeated-measures F-test –11.374 6.057 –23.2457 0.49772 0.09 Not shown Speed ACQ Overall effect Continuous variable ANOVA repeated-measures F-test F = 4.883 0.006 Not shown Speed ACQ TS_WT Continuous variable ANOVA repeated-measures F-test –4.547 1.354 –7.20084 –1.89316 0.004 Not shown Speed ACQ TS_TS EE-EGCG Continuous variable ANOVA repeated-measures F-test 1.725 1.457 –1.13072 4.58072 0.28 Not shown Speed ACQ WT EE-EGCG _WT Continuous variable ANOVA repeated-measures F-test –0.485 1.216 –2.86836 1.89836 0.69 Fig. 1E Latency first entry REM TS_WT Continuous variable One-way ANOVA 25.284 7.295 10.9858 39.5822 0.004 Fig. 1E Latency first entry REM TS EE-EGCG _TS Continuous variable One-way ANOVA –16.075 7.850 –31.461 –0.689 0.096 Fig. 1F Latency first entry REM Overall effect Continuous variable One-way ANOVA F = 6.159 0.002 Fig. 1F Gallagher index REM TS_WT Continuous variable One-way ANOVA 17.991 7.041 4.19064 31.79136 0.03 Fig. 1F Gallagher distance REM TS_WT Continuous variable One-way ANOVA 1272.4 455.5 –437.28 2165.18 0.025 Fig. 1F Gallagher index REM TS EE-EGCG _TS Continuous variable One-way ANOVA –11.944 7.577 –26.7949 2.90692 0.14 Fig. 1G Latency REV Overall effect Continuous variable Tobit model χ2(3) = 26.59 <0.001 Fig. 1G Latency REV TS_WT Continuous variable Tobit model 35.093 9.16 17.124 53.063 <0.001 Fig. 1G Latency REV TS EE-EGCG _TS Continuous variable Tobit model –14.865 9.28 –33.073 3.342 0.060 Fig. 1H Gallagher distance REV Overall effect Continuous variable ANOVA repeated-measures F-test F = 7.694 <0.001 Fig. 1H Gallagher distance REV TS_WT Continuous variable ANOVA repeated-measures F-test 2114.3 412 1306.78 2921.82 <0.001 Fig. 1H Gallagher distance REV TS EE-EGCG _TS Continuous variable ANOVA repeated-measures F-test –869 443.3 –1737.87 –0.132 0.059 Not shown Gallagher index REV Overall effect Continuous variable ANOVA repeated-measures F-test F = 11.714 <0.001 Not shown Gallagher index REV TS_WT Continuous variable ANOVA repeated-measures F-test 20.8045 4.3102 12.35651 29.25249 <0.001 Not shown Gallagher index REV TS EE-EGCG _TS Continuous variable ANOVA repeated-measures F-test –6.3488 4.638 –15.4393 2.74168 0.2 Fig. 1H Thigmotaxis REV Overall effect Continuous variable ANOVA repeated-measures F-test F = 10.105 <0.001 Fig. 1H Thigmotaxis REV TS_WT Continuous variable ANOVA repeated-measures F-test 21.393 6.268 9.10772 33.67828 0.001 Fig. 1H Thigmotaxis REV TS EE-EGCG _TS Continuous variable ANOVA repeated-measures F-test –9.867 6.745 –23.0872 3.3532 0.14 Not shown Speed REV Overall effect Continuous variable ANOVA repeated-measures F-test F = 2.607 0.067 - Table 2.
Multivariate within-group variances (sum over squared distances from group barycenter divided by group size, scaled by number of variables) before and after acquisition, and reversal sessions.
Figure Session WT WT-EE-EGCG TS TS-EE-EGCG Fig. 2E ACQ1 0.42 0.89 0.24 0.53 Fig. 2E ACQ5 1.62 1.55 0.87 0.95 Fig. 3E REV1 1.52 1.12 0.56 1.22 Fig. 3E REV3 2.68 1.84 1.42 2.26 Figure Variable Phase Contrast Pseudo-t p-value Fig. 2E PC1 ACQ1 TS_WT 3.67 <0.001 Fig. 2E PC1 ACQ1 TS_WT EE-EGCG 3.57 <0.001 Fig. 2E PC1 ACQ1 TS_TS EE-EGCG 3.05 0.004 Fig. 2E PC1 ACQ1 TS EE-EGCG_WT 0.28 0.39 Fig. 2E PC1 ACQ1 TS EE-EGCG_WT EE-EGCG 0.54 0.71 Fig. 2E PC1 ACQ1 WT_WT EE-EGCG 0.8 0.89 Fig. 2E PC1 ACQ5 TS_WT 6.81 <0.001 Fig. 2E PC1 ACQ5 TS_WT EE-EGCG 6.72 <0.001 Fig. 2E PC1 ACQ5 TS_TS EE-EGCG 1.85 0.045 Fig. 2E PC1 ACQ5 TS EE-EGCG_WT 5.2 9.99 Fig. 2E PC1 ACQ5 TS EE-EGCG_WT EE-EGCG 5.06 <0.001 Fig. 2E PC1 ACQ5 WT_WT EE-EGCG 0.27 0.39 Fig. 3E PC1 REV1 TS_WT 4.60 <0.001 Fig. 3E PC1 REV1 TS_WT EE-EGCG 5.60 <0.001 Fig. 3E PC1 REV1 TS_TS EE-EGCG 2.59 0.01 Fig. 3E PC1 REV1 TS EE-EGCG_WT 2.44 0.01 Fig. 3E PC1 REV1 TS EE-EGCG_WT EE-EGCG 3.07 0.005 Fig. 3E PC1 REV1 WT_WT EE-EGCG 0.23 0.58 Fig. 3E PC1 REV3 TS_WT 5.17 <0.001 Fig. 3E PC1 REV3 TS_WT EE-EGCG 6.19 <0.001 Fig. 3E PC1 REV3 TS_TS EE-EGCG 2.32 0.02 Fig. 3E PC1 REV3 TS EE-EGCG_WT 2.66 0.01 Fig. 3E PC1 REV3 TS EE-EGCG_WT EE-EGCG 3.21 0.004 Fig. 3E PC1 REV3 WT_WT EE-EGCG 0.21 0.58 95% confidence interval Figure Contrast Data structure Type of test Estimated mean difference Lower Higher p-value Fig. 4 TS _WT Continuous variable ANOVA –21.728 –42.94 –0.511 0.083 Fig. 4 TS EE-EGCG_TS Continuous variable ANOVA 27.224 5.311 49.136 0.044 Fig. 4 TS EE-EGCG _WT Continuous variable ANOVA 5.496 –15.721 26.713 0.616 Fig. 4 WT EE-EGCG_WT Continuous variable ANOVA –42.928 –64.145 –21.711 0.001 95% confidence interval Figure Region Contrast Data structure Type of test Estimated mean difference Lower Higher p-value Fig. 5A DG TS _WT Continuous variable Mixed-model F-test –0.192 –0.383 –0.002 0.048 Fig. 5B CA1 TS _WT Continuous variable Mixed-model F-test –0.175 –0.282 –0.069 <0.001 Fig. 5B CA1 TS EE-EGCG_TS Continuous variable Mixed-model F-test 0.105 0.001 0.209 0.047 Fig. 5B CA1 WT EE-EGCG_WT Continuous variable Mixed-model F-test –0.129 –0.234 –0.025 0.01 95% confidence interval Figure Variable Region Contrast Data structure Type of test Estimated mean difference Lower Higher p-value Not shown VGLUT1 puncta density DG TS _WT Continuous variable Mixed-model F-test 0.077 0.018 0.136 0.006 Not shown VGLUT1 puncta size DG TS _WT Continuous variable Mixed-model F-test –0.06 –0.098 –0.022 0.001 Fig. 6A VGLUT1/VGAT puncta density DG TS _WT Continuous variable Mixed-model F-test 0.308 0.151 0.465 <0.001 Fig. 6B VGLUT1/VGAT % of area DG TS _WT Continuous variable Mixed-model F-test 0.077 –0.03 0.183 0.222 Not shown VGLUT1puncta density DG TS EE-EGCG_TS Continuous variable Mixed-model F-test –0.073 –0.132 –0.014 0.01 Not shown VGLUT1 puncta size DG TS EE-EGCG_TS Continuous variable Mixed-model F-test 0.06 0.022 0.098 0.001 Fig. 6A VGLUT1/VGAT puncta density DG TS EE-EGCG_TS Continuous variable Mixed-model F-test –0.294 –0.452 –0.137 <0.001 Not shown VGLUT1 puncta density CA1 TS _WT Continuous variable Mixed-model F-test 0.071 0.008 0.134 0.022 Not shown VGLUT1 puncta size CA1 TS _WT Continuous variable Mixed-model F-test –0.093 –0.145 –0.041 0.043 Not shown VGAT puncta size CA1 TS _WT Continuous variable Mixed-model F-test 0.054 0.001 0.106 0.043 Fig. 6C VGLUT1/VGAT puncta density CA1 TS _WT Continuous variable Mixed-model F-test 0.295 0.098 0.493 0.001 Fig. 6D VGLUT1/VGAT % of area CA1 TS _WT Continuous variable Mixed-model F-test –0.145 –0.275 –0.016 0.023 Fig. 6C VGLUT1/VGAT puncta density CA1 TS EE-EGCG_TS Continuous variable Mixed-model F-test –0.245 –0.442 –0.047 0.01 Not shown VGAT puncta size CA1 WT EE-EGCG_WT Continuous variable Mixed-model F-test 0.049 –0.003 0.102 0.07 Fig. 6D VGLUT1/VGAT % of area CA1 WT EE-EGCG_WT Continuous variable Mixed-model F-test –0.137 –0.267 –0.007 0.035
In this issue
