Temporal aspects of behavioral and dendritic changes following acute CORT treatment have not been examined yet. Therefore, in the present study, we explored the temporal effects of acute CORT or vehicle administration by measuring dendritic architecture and behavior with variable delays, i.e., 3, 6, 12, 20 days after the treatment (Figure 1A). When we first measured a temporal change of body weight following acute CORT treatment (10 mg/kg of body weight), a two-way repeated-measures ANOVA revealed significant main effect of time (F(3, 54) = 391.9, P < 0.0001) and of treatment (F(1, 54) = 5.54, P < 0.05) but not of interaction (F(3, 54) = 1.79, P > 0.15) (Figure 1B). Delayed significant difference in body weight by acute CORT treatment was revealed on day 20 (P < 0.05, post hoc Bonferroni test).
Dendritic architecture of BLA neurons
Total dendritic length (Figure 2A) and number of branch points (Figure 2B) of BLA neurons were measured at various time points following a treatment of vehicle or CORT. Seven animals were assigned to each group and 20 ~ 26 neurons (about 3 neurons per an animal) were used to measure dendritic architectures. Total dendritic length and number of branch point for each animal were averaged, from which the group averages were calculated.
Overall, hypertrophic alteration of dendritic arborization of BLA neurons upon acute CORT treatment was significantly delayed and reversible. In terms of dendritic length of BLA neurons, a two-way ANOVA revealed significant main effects of time (F(3, 48) = 40.98, P < 0.0001), of treatment (F(1, 48) = 25.39, P < 0.0001), and of their interaction (F(3, 48) = 33.55, P < 0.0001). Number of branch points of BLA neurons also showed significant main effects of time (F(3, 48) = 16.43, P < 0.0001), and of their interaction (F(3, 48) = 7.39, P = 0.0004), though main effect of treatment did not reach statistical significance (F(1, 48) = 2.167, P > 0.1).
We analyzed the differences in detail between vehicle-treated and CORT-treated groups at each time point by post-hoc Bonferroni test following a two-way ANOVA. On day 3, no significant difference between vehicle- and CORT-treated groups was found either in total dendritic length (vehicle: 1028.2 ± 47.1 μm, CORT: 870.8 ± 48.9 μm, P > 0.05, Figures 2A-3d) or in number of branch points (vehicle: 10.52 ± 0.54, CORT: 9.38 ± 0.59, P > 0.05, Figures 2B-3d). To investigate the effects of acute CORT treatment on dendritic architecture in greater detail, segmental analysis was performed to track changes in dendritic length as a function of radial distance from the soma (segmental distance: 20 μm; Figure 2C). This analysis further confirmed that all the aspects of dendritic morphology were comparable between vehicle- and CORT-treated groups on day 3 (Figures 2C-3d).
On day 6, CORT-treated group showed slightly increased total dendritic length compared to vehicle-treated group, although the difference was not statistically significant (16.3% increase; vehicle: 972.9 ± 38.3 μm, CORT: 1131.9 ± 55.5 μm; P > 0.05; Figure 2A-6d). Number of branch points of CORT-treated group was comparable to that of vehicle-treated group (vehicle: 11.19 ± 0.46, CORT: 11.31 ± 0.56, P > 0.05, Figure 2B-6d). However, segmental analysis revealed significant dendritic expansion at 80 μm (P < 0.01) and 100 μm (P < 0.05) from the soma in CORT-treated group (Figure 2C-6d). Thus, hypertrophic effect of acute CORT began appearing 6 days after the treatment.
On day 12, CORT treatment dramatically increased total apical dendritic length (77.7% increase; vehicle: 1079.9 ± 76.2 μm, CORT: 1918.9 ± 55.2 μm, P < 0.0001, Figure 2A-12d) and total number of branch points (35.7% increase; vehicle: 12.17 ± 0.58, CORT: 16.52 ± 0.81, P < 0.0001, Figure 2B-12d). Segmental analysis also showed dendritic hypertrophy in CORT-treated group at 40 μm (P < 0.01) and within a distance of 60 ~ 160 μm (P < 0.001) from the soma (Figure 2C-12d).
On day 20, interestingly, all the dendritic measurements revealed that the dramatic hypertrophic effect was returned to the level of vehicle-treated groups, which indicates the hypertrophic effect of CORT on dendrites of BLA neurons was reversible. Total dendritic length (vehicle: 1060.9 ± 63.2 μm; CORT: 1001.6 ± 45.4 μm; P > 0.5; Figure 2A-20d) and number of branch points (vehicle: 10.71 ± 0.99; CORT: 10.10 ± 0.52; p > 0.5; Figure 2B-20d) returned to the vehicle-treated level, and any significant changes in detailed dendritic architecture were not observed (Figure 2C-20d).
Representative camera lucida drawings of Golgi-impregnated BLA pyramidal neurons for vehicle- or CORT- treated groups with various delay periods after the treatment are illustrated in Figure 2D.
Anxiety
Stressful events induce glucocorticoid release and evoke anxiety. Amygdalar hyperactivity is known to be accompanied by anxiety-like behavior [17–20]. Therefore, we examined whether acute treatment with single dose of CORT leads to a temporal change in anxiety behavior which is paralleled by the delayed yet reversible temporal change in dendritic architecture of BLA. Anxieties of animals 3, 6, 12 and 20 days following an acute vehicle or CORT treatment were measured in terms of reduction in open arm exploration in the elevated plus maze (Figure 3). Ten to eleven animals were assigned to each condition, i.e., total 84 animals were used for the anxiety measurement.
As shown in Figure 3A, a two-way ANOVA revealed significant main effects of time (F(3, 76) = 3.89, P < 0.05) and of interaction (F(3, 76) = 3.17, P < 0.05) on open arm duration. However, main effect of treatment did not reach statistical significance (F(1, 76) = 0.53, P > 0.5). In terms of open arm entry, only main effect of time was significant (F(3, 76) = 6.76, P = 0.0004, Figure 3B). Post hoc Bonferroni test showed that percent open arm duration (P > 0.05, Figure 3A-3d, -6d) and number of total open arm entry (P > 0.05, Figure 3B-3d, -6d) were comparable between vehicle- and CORT-treated groups 3 days and 6 days after the treatments. Meanwhile, on day 12, CORT-treated group exhibited a significantly greater degree of anxiety compared with vehicle-treated group. This elevated anxiety level was manifested as a significant reduction in percentage of open arm duration (48.9% reduction; vehicle: 29.16 ± 3.08%, CORT: 14.89 ± 2.50%; P < 0.05; Figure 3A-12d) and number of open arm entry (55.2% reduction; vehicle: 23.90 ± 3.80, CORT: 10.70 ± 1.95; P < 0.05; Figure 3B-12d).
Interestingly, the enhanced anxiety was not persistent. On day 20, anxiety level of CORT-treated group returned to that of vehicle-treated group; both open arm duration (vehicle: 30.76 ± 3.35%, CORT: 32.15 ± 2.30%; P > 0.05; Figure 3A-20d) and open arm entry (vehicle: 30.36 ± 3.80, CORT: 30.00 ± 4.71; P > 0.05; Figure 3B-20d) of CORT-treated group were comparable to those of vehicle-treated group.
As shown in Figure 3C, a two-way ANOVA revealed no significant main effects of interaction (F(3, 76) = 1.62, P > 0.1), of treatment (F(1, 76) = 0.057, P < 0.8) and of time (F(3, 76) = 0.79, P > 0.5) on total moving distance, which implicates that the higher anxiety was not accompanied by difference in locomotion activity. Overall, acute CORT induced dendritic hypertrophy of BLA spiny neurons, which was paralleled by heightened anxiety, both peaked 12 days after the treatment. Therefore, dendritic arborization of BLA neurons was closely related with anxiety level.
Dendritic architecture of mPFC neurons
In addition to the BLA, the mPFC has been known to be one of the stress vulnerable brain areas. The mPFC in rodents includes the dorsal anterior cingulate cortex (ACd), prelimbic PFC (PL) and infralimbic PFC (IL). Apical dendritic arborization of pyramidal neurons located in the PL and IL was measured across days in terms of total dendritic length (Figure 4A) and number of branch points (Figure 4B). A detailed segmental analysis was also performed using radial distance from the soma (Segmental distance, 40 μm; Figure 4C). Seven animals were assigned to each condition and 20 ~ 27 neurons (about 3 neurons per an animal) were used for each measurement.
Overall, delayed and reversible atrophic alteration of dendritic arborization of mPFC neurons upon acute CORT treatment was observed. In terms of total apical dendritic length, a two-way ANOVA revealed significant main effects of time (F(3,48 = 4.63, P = 0.0064), of treatment (F(1, 48) = 14.44, P = 0.0004), and of their interaction (F(3, 48) = 3.03, P = 0.0381). Number of branch points of mPFC neurons also showed significant main effects of time (F(3, 48) = 9.77, P < 0.0001), of treatment (F(1, 48) = 30.57, P < 0.0001) and of their interaction (F(3, 48) = 11.96, P < 0.0001). The difference between vehicle-treated and CORT-treated groups at each time point was further analyzed in detail by post-hoc Bonferroni test following a two-way ANOVA.
On day 3, there was a trend of decrease in dendritic arborization in the CORT-treated group compared with the vehicle-treated group. Although a decrease of total dendritic length in the CORT-treated group did not reach statistical significance (12.4% decrease; vehicle: 1739.9 ± 74.3 μm, CORT: 1523.6 ± 89.7 μm, P > 0.05, Figure 4A-3d), number of branch points significantly decreased in the CORT-treated group (15.8% decrease; vehicle: 15.48 ± 0.52, CORT: 12.72 ± 0.63, P < 0.05, Figure 4B-3d). Segmental analysis (segmental distance: 40 μm) also revealed that dendritic atrophy in the CORT-treated group began to be apparent at 160 μm from soma on day 3 (P < 0.05, Figure 4C-3d).
On day 6, CORT-treated group showed a significant decrease in total dendritic length (28.2% decrease; vehicle: 1824.9 ± 82.7 μm, CORT: 1309.8 ± 57.7 μm, P < 0.001, Figure 4A-6d) as well as in number of branch points (39.9% decrease; vehicle: 16.18 ± 0.55, CORT: 9.71 ± 0.44, P < 0.001, Figure 4B-6d). Segmental analysis revealed significant dendritic atrophy at 120 μm (P < 0.001) and 160 μm (P < 0.05) from soma in CORT-treated group (Figure 4C-6d).
However, 12 days after an acute CORT treatment, atrophic effects of CORT on dendrite morphology dramatically disappeared, which maintained thereafter; total dendritic length (vehicle: 1823.8 ± 66.8 μm, CORT: 1743.4 ± 65.3 μm, P > 0.05, Figure 4A-12d) and number of branch points (vehicle: 15.86 ± 0.49, CORT: 15.77 ± 0.80, P > 0.05, Figure 4B-12d) returned to the vehicle-treated level and no significant changes in detailed dendritic architecture were observed (Figure 4C-12d). Also, both total apical dendritic length (vehicle: 1877.9 ± 93.6 μm; CORT: 1796.4 ± 118.6 μm; P > 0.5; Figure 4A-20d) and total number of branch points (vehicle: 15.71 ± 0.54; CORT: 15.47 ± 0.812; p > 0.5; Figure 4B-20d) were comparable. The segmental analysis also confirmed the disappearance of CORT effect on day 20 (Figure 4C-20d). Therefore, the atrophic effect of CORT on dendrite of mPFC neurons is reversible. Figure 4D shows camera lucida drawings of representative Golgi-impregnated mPFC pyramidal neurons from vehicle- and CORT-treated animals.
Working memory
Apical dendritic atrophy and spine loss in mPFC are structural changes that result from experiencing traumatic stress [12, 16, 28, 29] and these changes may be associated with altered emotionality, impaired working memory, and dysfunctional regulation of stress hormone homeostasis [11, 30–33]. In the present study, we found that alteration of the anxiety level after an acute CORT treatment follows the time course change of dendritic architecture of BLA neurons but not mPFC neurons, indicating that enhanced anxiety is more likely associated with BLA hypertrophy rather than mPFC atrophy. Therefore, we investigated whether mPFC atrophy upon acute CORT treatment is rather accompanied by working memory deficit.
We assessed the temporal effect of acute CORT treatment on working memory performance using the Y-maze test. Percentage alternation calculated as the ratio of actual to possible alternation was considered as a parameter for working memory-related behaviors. Percent alternations of animals in the Y-maze 3, 6, 12 and 20 days following an acute vehicle or CORT treatment were measured (Figure 5). Ten animals were assigned to each condition and total 80 animals were used for the anxiety measurement. As shown in Figure 5, a two-way ANOVA revealed a significant main effect of interaction (F(3, 72) = 2.94, P < 0.05). However, main effect of treatment (F(1, 72) = 0.47, P > 0.5) and of time (F(3, 72) = 1.64, P > 0.1) did not reach statistical significance. Post hoc Bonferroni test showed that working memory impairment only in the CORT-treated animals on day 6 was manifested as a significant reduction in percentage of spontaneous alternations (17.3% reduction; vehicle: 78.59 ± 1.93%, CORT: 64.95 ± 2.68%; P < 0.05; Figure 5-6d).
Regardless of CORT treatment, however, percentage of alternation was comparable to the vehicle-treated animals on days 3, 12 and 20 (P > 0.05, Figure 5-3d, -12d, -20d). Our result clearly shows coincident occurrence of the working memory impairment and the dendrite atrophy of mPFC neurons in CORT-treated group 6 day after the treatment. There was a strong trend that high and low level of dendritic arborizations of mPFC neurons is accompanied by high and low performance of working memory, respectively.