Fluoxetine treatment cessation increases ethanol self-administration during reinstatement
As expected, a significant increase in weekly ethanol self-administration during the re-exposure period (reinstatement) was detected in the rats that were previously treated with fluoxetine during the ethanol deprivation period (Fig. 1B). Repeated measures ANOVA indicated an overall treatment effect on ethanol self-administration for 3 weeks (F1,72=26.66, p<0.0001), with fluoxetine-treated rats having higher consumption of ethanol during reinstatement (post hoc comparisons: *p<0.05 vs. ethanol-exposed rats treated with vehicle). Interaction between time and treatment was found when baseline period was introduced in the statistical analysis (F3,96=3.56, p<0.02), suggesting that fluoxetine treatment cessation modified ethanol self-administration during reinstatement compared to ethanol self-administration baseline.
Ethanol drinking reinstatement increases IBA-1 immunoreactivity in the brain
To analyze the effects of the abrupt cessation of fluoxetine treatment during ethanol abstinence and ethanol drinking reinstatement on the presence of microglia in the rat brain, we quantified the IBA-1+ cell number and IBA-1 immunoreactivity (intensity) in the PrL, Str, BLA, and the hippocampal areas CA1 and CA3, and the dentate gyrus.
Alcohol, but not saccharine drinking induced significant overall effects on the IBA-1 immunoreactivity, but not IBA-1+ cell number, in the PrL (F1,20=12.13, p=0.0027), Str (F1,20=29.88, p<0.0001) and BLA (F1,20=46.08, p<0.0001), showing a higher intensity in both ethanol-exposed rats (post hoc comparisons: **/***p<0.01/0.001 vs. saccharine-exposed rats) and ethanol-exposed rats treated with fluoxetine (post hoc comparisons: $$$p<0.001 vs. saccharine-exposed rats treated with fluoxetine, Figs. 2A-I). Treatment overall effect and interaction between drinking and treatment were not detected in the three brain regions.
Drinking induced a significant overall effect on IBA-1+ cell number in the dorsal hippocampus (F1,20=15.03, p=0.001), showing a specific increase in the dentate gyrus (overall effect: F1,20=154.4, p<0.0001), but not CA3 and CA1 areas, of both ethanol-exposed rats (post hoc comparisons: ***p<0.001 vs. saccharine-exposed rats) and ethanol-exposed rats treated with fluoxetine (post hoc comparisons: $$$p<0.001 vs. saccharine-exposed rats treated with fluoxetine, Figs. 3A-C). Drinking also resulted in an overall effect on IBA-1 immunoreactivity in the dorsal hippocampus (F1,20=31.36, p<0.0001), including dentate gyrus (F1,20=21.41, p<0.001), CA3 (F1,20=22.15, p<0.001) and CA1 (F1,20=32.67, p<0.0001). As a consequence, ethanol increased IBA-1 immunoreactivity in these hippocampal regions of both ethanol-exposed rats (post hoc comparisons: **/***p<0.01/0.001 vs. saccharine-exposed rats) and ethanol-exposed rats treated with fluoxetine (post hoc comparisons: $/$$p<0.05/0.01 vs. saccharine-exposed rats treated with fluoxetine, Figs. 3D-F). Treatment overall effect and interaction between drinking and treatment were not detected in dorsal hippocampus. Representative images of the immunohistochemical expression of IBA-1 and apparent density of IBA-1+ cells are shown in Figs. 3G-I.
Fluoxetine treatment cessation before ethanol drinking reinstatement modifies microglial morphology in a brain region-specific manner
We assessed whether increases in IBA-1 immunoreactivity induced by ethanol are associated with changes in morphometric parameters of microglial cells expressing IBA-1 in the PrL, Str, BLA and the hippocampal CA1 region. We also evaluated whether fluoxetine treatment cessation modifies the putative ethanol effect on microglial morphology.
Analysis of interaction and main overall effects of drinking and treatment
Significant overall effects of drinking (saccharine vs. ethanol) on fractal dimension (spatial complexity), lacunarity (heterogeneity), roughness (process surface and branching), and cell area, perimeter and circularity were detected in the microglia of the PrL (Table 2). We also observed significant overall effects of drinking on most morphometric parameters, excepting lacunarity, convex hull span ratio (CHSR) and the maximum/minimum convex hull radii (RCHR), in the microglia of the Str. Significant effects of drinking on lacunarity and density (process shortening/thickening) were only observed in the microglia of the BLA. Finally, significant overall effects of drinking on most morphometric parameters, excepting lacunarity, density, cell circularity, convex hull circularity (CHC), bounding circle diameter (BCD) and maximum span across the convex hull (MSACH), were observed in the microglia of the hippocampal CA1 region (Table 2).
Significant overall effects of treatment (vehicle vs. fluoxetine) on BCD, MSACH, and the mean radius (MR) were detected in the microglia of the PrL, and no treatment effect on the morphometric parameters in the microglia of the BLA was found (Table 2). Prominently, we found significant overall effects of treatment on fractal dimension, roughness, cell perimeter and circularity, and convex hull area and perimeter in the microglia of the Str. Significant overall treatment effects on fractal dimension, cell circularity, convex hull area, perimeter and circularity, density, BCD, MSACH, and MR were also found in the microglia of the hippocampal CA1 region (Table 2).
Interaction between factors (drinking vs. treatment) was only observed in roughness in the microglia of the PrL, suggesting a specific increase in process surface and branching in ethanol-exposed rats that were previously treated with fluoxetine. Regarding the striatum, interaction was mainly detected in microglial cell circularity (Table 2), suggesting that fluoxetine decreases the proportion between microglial cell area and perimeter in a drinking-dependent manner. In the microglia of CA1 region, interaction was found in fractal dimension, lacunarity and density, with fluoxetine decreasing spatial complexity and increasing heterogeneity and process length/thinning in a drinking-dependent manner. Interaction was not observed when morphometric parameters of microglia were analyzed in BLA (Table 2).
Simple effect analysis of fractal dimension, lacunarity, and cell area, perimeter and circularity
Following two-way ANOVA of main factors, Tukey’s post hoc for multiple comparisons were conducted when appropriate. Ethanol increased fractal dimension (spatial comlexity) in the Str of vehicle-treated rats (*p<0.05, Fig. 4B), as well as fractal dimension in the PrL, Str, BLA and CA1 of rats treated with fluoxetine ($/$$p<0.05/0.01, Figs. 4A-D). Fractal dimension is increased in the Str and decreased in the CA1 of saccharine-exposed rats that were previously treated with fluoxetine (*p<0.05, Figs. 4B, 4D). Ethanol decreased lacunarity (heterogeneity) in the CA1 of vehicle-treated rats (*p<0.05, Fig. 4H), as well as lacunarity in the PrL and CA1 of fluoxetine-treated rats ($$p<0.01, Figs. 4E, 4H). Fluoxetine increased lacunarity in the PrL and CA1 of saccharine-exposed rats (*/***p<0.05/0.001, Figs. 4E, 4H). Microglial cell area (process branching and/or soma enlargement) was increased by ethanol in the Str and CA1 of vehicle-treated rats (**/***p<0.01/0.001, Figs. 4J, 4L), as well as in the PrL, Str, BLA and CA1 of fluoxetine-treated rats ($/$$/$$$p<0.05/0.01/0.001, Figs. 4I-L). Microglial cell area was decreased in the PrL and increased in the Str of saccharine-exposed rats that were previously treated with fluoxetine (*p<0.05, Figs. 4I, 4J). Interestingly, fluoxetine increased microglial cell area in the CA1 of ethanol-exposed rats (#p<0.05, Fig. 4L). Ethanol increased microglial cell perimeter in the Str and CA1 of vehicle-treated rats (*/**p<0.05/0.01, Figs. 4N, 4P), as well as cell perimeter in the PrL and CA1 of fluoxetine-treated rats ($/$$$p<0.05/0.001, Figs. 4M, 4P). Ethanol decreased microglial cell circularity in the Str of vehicle-treated rats (***p<0.001, Fig. 4R), as well as cell circularity in the PrL of fluoxetine-treated rats ($$p<0.01, Fig. 4Q). Fluoxetine also decreased cell circularity in the Str and CA1 of saccharine-exposed rats (**p<0.01, Figs. 4R, 4T). Interestingly, fluoxetine specifically decreased cell circularity in the CA1 of ethanol-exposed rats (#p<0.05, Fig. 4T).
Simple effect analysis of density, roughness, and convex hull area, perimeter and circularity
Following two-way ANOVA of main factors, Tukey’s post hoc for multiple comparisons were conducted when appropriate. Ethanol increased convex hull (CH) area in the Str and CA1 of vehicle-treated rats (**p<0.01, Figs. 5B, 5D), as well as CH area in the PrL and CA1 of fluoxetine-treated rats ($p<0.05, Figs. 4A, 5D). Fluoxetine decreased CH area in the PrL and BLA, but increased CH area in the Str and CA1 of saccharine-exposed rats (*/**p<0.05/0.01, Figs. 5A-D). Ethanol increased density in the BLA of vehicle-treated rats (**p<0.01, Fig. 5G), as well as density in the Str and CA1 of fluoxetine-treated rats ($/$$p<0.05/0.01, Figs. 5F, 5H). In contrast, fluoxetine specifically decreased density in the CA1 of saccharine-exposed rats (***p<0.001, Fig. 5H). Ethanol increased CH perimeter in the Str and CA1 of vehicle-treated rats (*/**p<0.05/0.01, Figs. 5J, 5L), as well as CH perimeter in the PrL of fluoxetine-treated rats ($p<0.05, Fig. 5I). Similar to CH area, fluoxetine decreased CH perimeter in the PrL and BLA, but increased CH perimeter in the Str and CA1 of saccharine-exposed rats (*/**p<0.05/0.01, Figs. 5I-L). Ethanol increased roughness in the Str and CA1 of vehicle-treated rats (*/**p<0.05/0.01, Figs. 5N, 5P), as well as roughness in the PrL, BLA and CA1 of fluoxetine-treated rats ($/$$/$$$p<0.05/0.01/0.001, Fig. 5M, 5O, 5P). Fluoxetine specifically increased roughness in the Str of saccharine-exposed rats (*p<0.05, Fig. 5N). Fluoxetine also increased cell circularity in the BLA and CA1 of ethanol-treated rats (#/###p<0.05/0.001, Figs. 5S, 5T), as well as cell circularity in the CA1 of the saccharine-treated rats (***p<0.001, Fig. 5T).
Simple effect analysis of convex hull span ratio, bounding circle diameter, maximum span across the convex hull, the ratio maximum/minimum convex hull radii, and the mean radius
Following two-way ANOVA of main factors, Tukey’s post hoc for multiple comparisons were conducted when appropriate. Ethanol specifically decreased CH span ratio in the CA1 of both vehicle- and fluoxetine-treated rats (***/$$$p<0.001, Fig. 6D). Ethanol increased bounding circle diameter (BCD) and maximum span across the convex hull (MSACH) in the Str of vehicle-treated rats (*/**p<0.05/0.01, Figs. 6F, 6J). In contrast, fluoxetine decreased BCD and MSACH in the PrL, but increased these parameters in the Str and CA1 of saccharine-exposed rats (*p<0.05, Figs. 6E, 6F, 6H, 6I, 6L). Similar to CH span ratio, ethanol decreased the ratio maximum/minimum CH radii (RCHR) in the CA1 of both vehicle- and fluoxetine-treated rats (***/$$p<0.001/0.01, Fig. 6P), as well as RCHR in the BLA of fluoxetine-treated rats ($p<0.05, Fig. 6O). Ethanol increased the mean radius (MR) in the Str and CA1 of vehicle-treated rats (*/***p<0.05/0.001, Figs. 6R, 6T), as well as MR in the PrL of fluoxetine-treated rats ($p<0.05, Fig. 6Q). In contrast, fluoxetine decreased MR in the PrL, but increased MR in the Str and CA1 of saccharine-exposed rats (*p<0.05, Figs. 6Q, 6R, 6T).
Microglia morphology correlates with fluoxetine-induced changes in inflammatory factors and TLR4 in a brain region-specific manner
Since alcohol induced specific activation of microglia, as measured by IBA1 immunoreactivity, we evaluated whether fluoxetine treatment cessation modifies mRNA expression of inflammatory factors (cytokines and chemokines) and TLR4 in the PrL, Str, BLA and the hippocampal CA1 region of those rats with alcohol drinking reinstatement. We also assessed whether specific morphometric features of microglia are tightly associated with changes in inflammatory factors induced by fluoxetine.
Fluoxetine reduced mRNA expression of the pro-inflammatory chemokine Cx3cl1 (fractalkine), and increased mRNA expression of the inflammatory cytokines IL1b and IL10 in the PrL of ethanol-exposed rats (*p<0.05, Fig. 7A). Fluoxetine only increased mRNA expression of TLR4 in the Str of ethanol-exposed rats (*p<0.05, Fig. 7B). Fluoxetine also reduced mRNA expression of the chemokine Cx3cl1, and increased mRNA expression of the anti-inflammatory cytokines IL4 in the BLA of ethanol-exposed rats (*p<0.05, Fig. 7C). Fluoxetine also reduced mRNA expression of the chemokine Cx3cl1, and increased mRNA expression of most of the remaining chemokines (Cxcl12, Ccl2) and cytokines (IL1β, IL6, IL10) analyzed, as well as BDNF and TLR4, in the dorsal hippocampus of ethanol-exposed rats (*p<0.05, Fig. 7D).
When we analyzed whether morphometric parameters of microglia correlated with these changes of inflammatory factors induced by fluoxetine, we specifically detected that fractal dimension (microglial spatial complexity) positively correlated with mRNA expression of IL10 (R= 0.67, p<0.009), and negatively correlated with mRNA expression of Cx3cl1 (R= -0.53, p<0.05) in the PrL (Fig. 7E). We also found that fluoxetine-induced increase in TLR4 expression specifically correlated with higher values of fractal dimension in the Str (R= 0.64, p<0.05, Fig. 7F). In the BLA (Fig. 7G), fractal dimension correlated with higher mRNA expression of IL4 (R= 0.85, p=0.0001), and lower mRNA expression of Cx3cl1 (R= -0.60, p<0.03) and Ccl2 (R= -0.65, p<0.02). Finally, in the dorsal hippocampus (Fig. 7H), microglial cell area (process branching and/or soma enlargement) positively correlated with fluoxetine-induced increases in mRNA expression of most factors analyzed (Cxcl2, Ccl2, IL1β, IL6, IL10, BDNF, TLR4: R> 0.55, p<0.04), and negatively correlated with fluoxetine-induced decreases in mRNA expression of Cx3cl1 (R= -0.66, p<0.01).