Males had higher inflammatory cytokines in the ventral hippocampus than females in adolescence. Adolescent males had higher IFN-γ, IL-1β, IL-4, IL-5, and IL-10 levels in the ventral hippocampus compared to adolescent females (p’s < 0.006; IFN-γ: sex by age interaction: F(2,48) = 5.865, p = 0.005, Ƞp2 = 0.196; IL-1β: sex by age interaction: F(2,47) = 5.557, p = 0.007, Ƞp2 = 0.191; IL-4: sex by age interaction: F(2,46) = 5.683, p = 0.006, Ƞp2 = 0.198; IL-5: sex by age interaction: F(2,47) = 6.229, p = 0.004, Ƞp2 = 0.21; IL-10: sex by age interaction: F(2,47) = 4.352, p = 0.019, Ƞp2 = 0.156). Adolescent males also had higher IFN-γ, IL-1β, IL-4, IL-5, IL-10, and IL-6 cytokine levels compared to young adult and middle-aged males (p’s < 0.011; IL-6: sex by age interaction: F(2,48) = 3.500, p = 0.038, Ƞp2 = 0.127) and adolescent and young adult females had higher IL-1β, IL-4, and IL-10 levels compared to middle-aged females (p’s < 0.046). See Fig. 1.
Middle-aged rats had higher CXCL1, IL-13, and TNF-α levels in the ventral hippocampus compared to young adults.
Regardless of sex, middle-aged rats had higher TNF-α levels in the ventral hippocampus compared to young adults (p = 0.004; main effect of age: F(2,48) = 5.96, p = 0.005, Ƞp2 = 0.199). There was a trend for middle-aged rats to have higher CXCL1 levels compared to adolescents (p = 0.063) and young adults (p = 0.059; main effect of age: F(2,49) = 3.12, p = 0.053, Ƞp2 = 0.113), regardless of sex. Both adolescents and middle-aged adults had higher IL-13 levels compared to young adults (p’s < 0.003; main effect of age: F(2,49) = 8.06, p = 0.001, Ƞp2 = 0.247). See Fig. 1.
Females had higher inflammatory cytokines in the basolateral amygdala than males in adulthood. Middle-aged rats had higher levels of TNF-α compared to other ages, regardless of sex. In contrast to the ventral hippocampus, females had higher levels of inflammation in the basolateral amygdala (BLA) in adulthood to middle-age than males, depending on the cytokine. Young adult females had higher levels of IFN-γ, IL-1β, IL-6, and IL-10 than young adult males [IFN-γ (p = 0.002; sex by age interaction: F(2,49) = 5.8003, p = 0.006, Ƞp2 = 0.191), IL-1β (p = 0.005; sex by age interaction: F(2,47) = 3.622, p = 0.034, Ƞp2 = 0.134), IL-6 (p = 0.001; sex by age interaction: F(2,46) = 8.44, p = 0.0008, Ƞp2 = 0.268), and IL-10 (p = 0.024; sex by age interaction: F(2,49) = 3.828, p = 0.029, Ƞp2 = 0.135), IL-5 (p = 0.0003; sex by age interaction approached significance: F(2,49) = 2.618, p = 0.08, Ƞp2 = 0.097), IL-13 (p = 0.002; sex by age interaction: F(2,49) = 5.482, p = 0.007, Ƞp2 = 0.183)]. There were no sex differences in these cytokines in adolescent or middle-aged rats (p’s > 0.424). Furthermore, young adult females had higher IFN-γ, IL-6, and IL-10 levels compared to all other groups (p’s < 0.048, although p = 0.09 compared to middle-aged females for IL-10). Middle-aged females had higher levels of IL-5 and IL-13 than middle-aged males and all other groups (p’s < 0.0002). Middle-aged males had higher CXCL1 levels in the BLA compared to middle-aged females and all other groups (p’s < 0.0002; sex by age interaction: F(2,44) = 10.03, p = 0.0003, Ƞp2 = 0.313). Regardless of sex, middle-aged rats had higher TNF-α levels in the BLA compared to adolescents and young adults (p’s < 0.0002; main effect of age: F(2,48) = 14.81, p = 0.00001, Ƞp2 = 0.382). See Fig. 2.
Negative correlations between basolateral amygdala and ventral hippocampal cytokines in young adulthood, sex difference in correlations between basolateral amygdala IL-6 and ventral hippocampus cytokines in adolescence. Correlations of cytokine and CXCL1 levels within the BLA and the ventral hippocampus were largely positive in all age groups, although correlations between regions were more negative in young adults, regardless of sex, compared to the other age groups. In adolescence, there was a sex difference in the correlations between BLA IL-6 and cytokines in the ventral hippocampus, with positive correlations in adolescent females compared to negative correlations in adolescent males (sex difference in BLA IL-6 and ventral hippocampal IFN-γ (z = 2.768, p = 0.003), IL-1β (z = 3.335, p < 0.001), IL-4 (z=-3.464, p < 0.001), IL-5 (z = 1.979, p = 0.024), IL-10 (z=-2.694, p = 0.004), IL-13 (z = 2.767, p = 0.003), TNF-α (z=-2.727, p = 0.003)). See supplementary Fig. S1.
We next correlated negative cognitive bias score with inflammatory markers in both the BLA or hippocampus but no correlations survived Bonferroni correction. See supplemental Table S1.
Neurogenesis decreased with age, regardless of sex. As expected, adolescents had higher DCX expression in the dorsal hippocampus compared to the ventral hippocampus (p = 0.0001) and higher DCX expression compared to the adult groups, regardless of sex (p’s < 0.00014). Moreover, young adults had higher DCX expression compared to middle-aged adults, regardless of sex (in dorsal, p = 0.0009, in ventral p = 0.03; region by age interaction: F(2,45) = 14.789, p = 0.00001, Ƞp2 = 0.397). See Fig. 3.
Negative cognitive bias was negatively associated with dorsal neurogenesis in young adult males. Dorsal DCX expression was negatively correlated with freezing (r= -0.787, p = 0.036) and negative cognitive bias score (r= -0.7643, p = 0.045) in young adult males only as there were no significant correlations with freezing or negative cognitive bias in any other group (p’s > 0.241). However, these correlations do not survive Bonferroni corrections. See supplemental Table S2.
Inflammation was positively associated ventral hippocampal neurogenesis in males but with dorsal hippocampal neurogenesis in females. In males, ventral hippocampal DCX was positively associated with ventral hippocampal IL-13 and TNF-α, but dorsal hippocampal DCX was negatively associated with BLA IL-1β (p’s < 0.043). However, only the correlation between ventral hippocampal DCX and IL-13 survived Bonferroni (p = 0.001). In females, dorsal hippocampal DCX was positively associated with ventral hippocampal IFN-γ, IL-5, IL-13, and CXCL1 in young adulthood (p’s < 0.03) but changed to positive correlations between the dorsal hippocampal DCX and BLA IL-4, IL-5, IL-10, and IL-13 in middle-age (p’s < 0.043). See supplemental Table S3.
Associations between cytokine/chemokine levels differ by age and sex after cognitive bias testing. Principal component analysis was used to identify clusters/pathways of interest or components [52–54]. The first two principal components accounted for 60.43% of the variance of all cytokine/chemokine and neurogenesis data. Component 1 accounts for 38.66% of the variance and is associated with cytokine/chemokine levels in the ventral HPC and hippocampal neurogenesis compared to cytokine/chemokine levels in the BLA. Component 2 accounts for 21.78% of the variance and is associated with all cytokines/chemokine levels in both regions. The loadings for PC1 and PC2 are shown in Table 1. An ANOVA on Principal Component 1 found that hippocampal inflammation and neurogenesis were higher in adolescents compared to the adult age groups (p’s < 0.00013), and IL-4, L-5, IL-13, TNF-α, and CXCL1 levels in the BLA were higher in middle-aged rats compared to the younger age groups (p’s < 0.00013) after cognitive bias testing (main effect of age: F(2,48) = 100.5, p < 0.000001, Ƞp2 = 0.807). A priori we expected sex differences and hippocampal inflammation and neurogenesis were elevated in adolescent males compared to adolescent females (p < 0.006; sex by age interaction: F(2,48) = 2.851, p < 0.068, Ƞp2 = 0.106) with no other sex differences seen. There was no significant main effect of sex (p > 0.102; see Fig. 4A). An ANOVA on Principal Component 2 found higher cytokine associations in young adult females compared to males (p = 0.005) after cognitive bias testing (sex by age interaction: F(2,48) = 8.578, p = 0.0007, Ƞp2 = 0.263). There were no other significant main or interaction effects (all p’s > 0.09). See Fig. 4B.
Table 1. Principal component loadings from DCX and cytokine/chemokine data.
|
|
PC1
|
PC2
|
Dorsal hippocampus
|
DCX
|
0.792*
|
-0.034
|
Ventral hippocampus
|
DCX
|
0.735*
|
-0.033
|
IFN-γ
|
0.846*
|
0.383*
|
IL-1β
|
0.910*
|
0.317*
|
IL-4
|
0.904*
|
0.312*
|
IL-5
|
0.593*
|
0.584*
|
IL-6
|
0.703*
|
0.338*
|
CXCL1
|
-0.173
|
0.381*
|
IL-10
|
0.937*
|
0.189
|
IL-13
|
0.424*
|
0.591*
|
TNF-α
|
-0.092
|
0.484*
|
Basolateral amygdala
|
IFN-γ
|
0.04
|
0.592*
|
IL-1β
|
-0.064
|
0.716*
|
IL-4
|
-0.761*
|
0.537*
|
IL-5
|
-0.743*
|
0.543*
|
IL-6
|
-0.201
|
0.567*
|
CXCL1
|
-0.567*
|
0.171
|
IL-10
|
-0.099
|
0.681*
|
IL-13
|
-0.678*
|
0.585*
|
TNF-α
|
-0.542*
|
0.433*
|
Significant loadings are indicated in bold. *indicates significance at p < 0.022. |