Effects of β-FNA on anxiety- and sickness-like behavior

Multiple behavioral traits were recorded and analyzed during the elevated plus maze test. Specifically, anxiety-like behavior was evaluated by examining time spent in closed arms, while sickness-like behavior was examined by evaluating time spent motionless, total distance moved, and time spent in a contracted body position. Two-way ANOVA results indicated a significant main effect of LPS (F1, 17 = 9.96, p < 0.01), no significant main effect of β-FNA (F1, 17 = 0.14, p = 0.71), and a significant interaction of main effects (F1, 17 = 72.42, p < 0.0001) on time spent in closed arms (Fig. 1A). Pairwise comparisons revealed that LPS mice spent significantly more time in the closed arms than saline (p < 0.0001) or β-FNA + LPS mice (p < 0.0001), and β-FNA + LPS mice spent significantly more time in closed arms than saline mice (p < 0.05). Notably, the β-FNA mice spent significantly more time in closed arms than saline mice (p < 0.0001) or β-FNA + LPS mice (p < 0.005). While β-FNA mice tended to spend less time in closed arms than LPS mice, the reduction in time was not significant (p = 0.07).

Fig. 1

Chronic effects of β-FNA on LPS-induced anxiety- and sickness-like behavior in male C57BL/6J mice. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. Data are presented as mean ± SEM. A Time in closed arms: Two-way ANOVA indicated significant main effect of LPS (p < 0.01), no significant main effect of β-FNA (p = 0.71), and a significant interaction of main effects (p < 0.0001) on time spent in the closed arms of the EPM (n = 5–6/group). Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05). B Time spent motionless: Two-way ANOVA indicated significant main effect of LPS (p < 0.005), no significant main effect of β-FNA (p = 0.12), and a significant interaction of main effects (p < 0.001) on time spent motionless in EPM (n = 5–8/group). C Total distance moved: Two-way ANOVA indicated significant main effect of LPS (p < 0.001), no significant main effect of β-FNA (p = 0.26), and a significant interaction of main effects (p < 0.05) on total distance moved in EPM (n = 6–8/group). D Time spent in a contracted position: Two-way ANOVA indicated significant main effects of LPS (p < 0.0001) and β-FNA (p < 0.0005), as well as a significant interaction of main effects (p < 0.0001) on time spent in a contracted position (n = 6–7/group). Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05)

Two-way ANOVA revealed a significant main effect of LPS (F1, 22 = 13.10, p < 0.005), no significant main effect of β-FNA (F1, 22 = 2.61, p = 0.12), and a significant interaction of main effects (F1, 22 = 15.57, p < 0.001) on time spent motionless in the elevated plus maze (Fig. 1B). Pairwise comparisons revealed that LPS mice spent significantly more time motionless than saline (p < 0.0001), β-FNA (p < 0.001), or β-FNA + LPS mice (p < 0.01), while time spent motionless between those three groups was similar (p = 0.10, p = 082, p = 0.20, respectively).

Two-way ANOVA also revealed a significant main effect of LPS (F1, 22 = 22.02, p < 0.001), no significant main effect of β-FNA (F1, 22 = 1.36, p = 0.26), and a significant interaction of main effects (F1, 22 = 7.90, p < 0.05) on total distance moved (Fig. 1C). Pairwise comparisons indicated that LPS mice covered significantly less distance than saline (p < 0.0001), β-FNA (p < 0.0005), or β-FNA + LPS (p < 0.05) mice. While distance covered was similar between saline and β-FNA mice (p = 0.25) and β-FNA and β-FNA + LPS mice (p = 0.21), β-FNA + LPS mice covered significantly less distance than saline mice (p < 0.05).

Finally, two-way ANOVA showed significant main effects of LPS (F1, 20 = 22.55, p < 0.001) and β-FNA (F1, 20 = 19.77, p < 0.0005), as well as a significant interaction of main effects (F1, 20 = 43.94, p < 0.0001) on time spent in a contracted position (Fig. 1D). LPS mice spent significantly more time in a contracted position than saline, β-FNA, or β-FNA + LPS mice (p < 0.0001). Additionally, time spent in a contracted position was similar between saline and β-FNA mice (p = 0.14), saline and β-FNA + LPS mice (p = 0.84), and β-FNA and β-FNA + LPS mice (p = 0.22).

In summary, chronic, continuous β-FNA treatment ameliorated LPS-induced anxiety- and sickness-like behaviors, as indicated by an increase in total distance moved, less time spent in closed arms, and less time spent motionless and in a contracted body position. β-FNA alone did not significantly affect sickness-like behaviors under non-inflammatory conditions; but it did increase anxiety-like behavior, as measured by increased time spent in closed arms.

Effects of β-FNA on depressive-like behavior

DBscorer software was used to automatically score key depressive-like traits in the FST, including percentage of time spent immobile and latency to immobility. Two-way ANOVA showed no significant main effect of LPS (F1, 19 = 0.55, p = 0.46), β-FNA (F1, 19 = 1.75, p = 0.20), or interaction of main effects (F1, 19 = 0.12, p = 0.74) on percentage of time spent immobile (Fig. 2A). Similarly, two-way ANOVA did not show a significant main effect of LPS (F1, 19 = 2.00, p = 0.17, β-FNA (F1, 19 = 1.94, p = 0.18), or interaction of main effects (F1, 19 = 3.31, p = 0.08) on latency to immobility (Fig. 2B). While β-FNA did not significantly reduce depressive-like behavior, control and β-FNA treatment groups tended to have increased latency to immobility compared to LPS mice.

Fig. 2

Chronic effects of β-FNA on LPS-induced depressive-like behavior in male C57BL/J6 mice. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. Endpoints measured included A percentage of time spent immobile and (B) latency to immobility. Data are presented as mean ± SEM. A Two-way ANOVA indicated no significant main effect of LPS (p = 0.46), β-FNA (p = 0.20), or interaction of main effects (p = 0.74) on percentage of time immobile (n = 5–7/group). B Two-way ANOVA indicated no significant main effect of LPS (p = 0.17), β-FNA (p = 0.18) or interaction of main effects (p = 0.08) on latency to immobility (n = 4–7/group). Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05)

Effects of β-FNA on CCL2 in the brain and spleen

CCL2 expression was quantified in frontal cortex, hippocampus, and spleen tissues (Fig. 3). Two-way ANOVA indicated that there was no significant main effect of LPS (F1, 19 = 0.98, p = 0.33) or β-FNA (F1, 19 = 2.74, p = 0.11) on CCL2 levels in the frontal cortex, but there was a significant interaction of main effects (F1, 19 = 67.23, p < 0.0001) (Fig. 3A). LPS mice had significantly higher levels of CCL2 in the frontal cortex than either saline (p < 0.0001) or β-FNA + LPS (p < 0.0001) mice, which were similar (p = 0.66). Additionally, β-FNA mice tended to have lower CCL2 levels in the frontal cortex than LPS mice (p = 0.06). Interestingly, β-FNA mice had significantly higher levels of CCL2 in the frontal cortex than either saline (p < 0.0005) or β-FNA + LPS mice (p < 0.0001).

Fig. 3

Chronic effects of β-FNA on LPS-induced elevations of CCL2 in male C57BL/J6 mice. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. CCL2 was measured via ELISA in (A) frontal cortex, (B) hippocampus, and (C) spleen tissues. Data are presented as mean ± SEM. A Two-ANOVA (n = 5–7/group) indicated that there was no significant main effect for LPS (p = 0.33) or β-FNA (p = 0.11) on CCL2 levels in the frontal cortex, but there was a significant interaction of main effects (p < 0.0001). B Two-way ANOVA revealed a significant main effect of LPS (p < 0.005), no significant main effect of β-FNA (p = 0.55), and a significant interaction of main effects (p < 0.005) on CCL2 levels in the hippocampus (n = 6–7/group). C Two-way ANOVA (n = 5–8/group) suggested a main effect for LPS (p < 0.0001) and β-FNA (p < 0.0001) on CCL2 levels in the spleen, but no interaction of main effects (p = 0.24). Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05)

Two-way ANOVA revealed a significant main effect of LPS (F1, 21 = 12.24, p < 0.005), no significant main effect of β-FNA (F1, 21 = 0.36, p = 0.55), and a significant interaction of main effects (F1, 21 = 11.50, p < 0.005) on CCL2 levels in the hippocampus (Fig. 3B). LPS mice had significantly higher levels of CCL2 in the hippocampus than saline (p < 0.0001), β-FNA (p < 0.01), or βFNA + LPS mice (p < 0.05). CCL2 levels in the hippocampus were similar between β-FNA and β-FNA + LPS mice (p = 0.94), but β-FNA + LPS mice had significantly higher CCL2 levels than saline mice (p < 0.05). While β-FNA mice tended towards higher levels of CCL2 in the hippocampus than saline mice, it fell short of significance (p = 0.06).

Two-way ANOVA indicated a main effect of LPS (F1, 21 = 38.03, p < 0.0001) and β-FNA (F1, 21 = 43.70, p < 0.0001) on CCL2 levels in the spleen, but no interaction of main effects (F1, 21 = 1.44, p = 0.24) (Fig. 3C). Pairwise comparisons revealed that LPS mice had significantly higher levels of CCL2 in the spleen than saline, β-FNA, or β-FNA + LPS mice (p < 0.0001). While saline and β-FNA + LPS mice had similar levels of CCL2 (p = 0.77), it is notable than β-FNA mice had significantly lower levels of CCL2 in the spleen than either saline (p < 0.001) or β-FNA + LPS mice (p < 0.005).

In summary, β-FNA abolished LPS-induced CCL2 elevations in the frontal cortex and spleen, while also significantly reducing CCL2 in the hippocampus. Under non-inflammatory conditions, β-FNA differentially affected select tissues by raising CCL2 levels in the frontal cortex; yet, reducing CCL2 levels in the spleen.

Effects of β-FNA on CXCL10 in the brain and spleen

Levels of the chemokine CXLC10 were also quantified in frontal cortex, hippocampus, and spleen (Fig. 4). Two-way ANOVA revealed a significant main effect of LPS (F1, 18 = 96.94, p < 0.0001), no significant main effect of β-FNA (F1, 18 = 0.51, p = 0.48), and a significant interaction of main effects (F1, 18 = 41.26, p < 0.0001) on CXCL10 levels in the frontal cortex (Fig. 4A). Pairwise comparisons revealed that LPS mice had significantly higher levels of CXCL10 in the frontal cortex than saline (p < 0.0001), β-FNA (p < 0.0001) or β-FNA + LPS (p < 0.005) mice. While saline (p < 0.0001) and β-FNA (p < 0.05) mice had significantly lower levels of CXCL10 in the frontal cortex than β-FNA + LPS mice, it is notable that β-FNA mice had significantly higher levels of CXCL10 than saline mice (p < 0.0001).

Fig. 4

Chronic effects of β-FNA on LPS-induced elevations in CXCL10 in male C57BL/6J mice. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. CXCL10 was measured via ELISA in (A) frontal cortex, (B) hippocampus, and (C) spleen tissues. Data are presented as mean ± SEM. A Two-way ANOVA revealed a significant main effect of LPS (p < 0.0001), no significant main effect of β-FNA (p = 0.48), and a significant interaction of main effects (p < 0.0001) on CXCL10 levels in the frontal cortex (n = 5–6/group). B Two-way ANOVA indicated that there were significant main effects of LPS (p < 0.0001) and β-FNA (p < 0.001), as well as an interaction of main effects (p < 0.005) on CXCL10 levels in the hippocampus (n = 5–8/group). C Two-way ANOVA (n = 6–7/group) revealed that there was a significant main effect for LPS (p < 0.0001) and β-FNA (p < 0.005) on CXCL10 levels in the spleen, but there was no significant interaction of main effects (p = 0.13). Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05)

Two-way ANOVA indicated that there were significant main effects of LPS (F1, 23 = 142.30, p < 0.001) and β-FNA (F1, 23 = 15.29, p < 0.001), as well as an interaction of main effects (F1, 23 = 11.90, p < 0.005) on CXCL10 levels in the hippocampus (Fig. 4B). LPS mice had significantly elevated levels of CXCL10 when compared to saline, β-FNA, and β-FNA + LPS mice (p < 0.0001). β-FNA + LPS mice had significantly higher CXCL10 levels than saline or β-FNA mice (p < 0.0001), while levels of CXCL10 were similar between saline and β-FNA mice (p = 0.72).

Two-way ANOVA revealed that there was a significant main effect for LPS (F1, 22 = 91.15, p < 0.0001) and β-FNA (F1, 22 = 12.06, p < 0.005) on CXCL10 levels in the spleen, but there was no significant interaction of main effects (F1, 22 = 2.49, p = 0.13) (Fig. 4C). LPS mice had significantly higher levels of CXCL10 than saline (p < 0.0001), β-FNA (p < 0.0001) or β-FNA + LPS mice (p < 0.005). CXCL10 levels were similar between saline and β-FNA mice (p = 0.19), and both saline (p < 0.0005) and β-FNA mice (p < 0.0001) had significantly lower levels of CXCL10 than β-FNA + LPS mice.

In conclusion, β-FNA attenuated LPS-induced elevations of CXCL10 in the frontal cortex, hippocampus, and spleen. Also, β-FNA prevented LPS-induced elevations of CCL2 in the brain and spleen. However, under non-inflammatory conditions, β-FNA raised CCL2 and CXCL10 levels in the frontal cortex and inhibited CCL2 expression in the spleen.

Correlation of inflammatory mediators in the brain and spleen with anxiety-like behavior

Linear regression analysis was performed to determine whether there were significant correlations between pro-inflammatory chemokine expression and anxiety-like behavior. CCL2 levels in the frontal cortex and hippocampus were positively correlated with time spent in closed arms (r2 = 0.56, F1, 16 = 20.81, p < 0.0005 and r2 = 040, F1, 13 = 8.67, p < 0.05, respectively; Fig. 5A). Furthermore, there was a significant positive correlation with CXCL10 expression in the frontal cortex and time spent in closed arms (r2 = 0.37, F1, 13 = 7.66, p < 0.05; Fig. 5B).

Fig. 5

Correlations between LPS-induced CXCL10 levels in male C57BL/6J mouse brains and anxiety-like behavior. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and delivered at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. A CCL2 and (B) CXCL10 were measured via ELISA in brain region and spleen homogenates. Data are presented as mean ± SEM. Linear regression analysis was used to assess frontal cortex, hippocampus, and spleen CCL2 and CXCL10 levels with EPM-time spent in closed arms. Linear regression statistics and symbols are provided in the figure, and only significant results are shown

Correlation of inflammatory mediators in the brain and spleen with sickness-like behavior

Linear regression analysis demonstrated that there were significant correlations between tissue-specific chemokine levels and sickness-like behaviors. Specifically, CCL2 levels in the frontal cortex, hippocampus, and spleen were each positively correlated with time spent motionless in the elevated plus maze (r2 = 0.41, F1, 19 = 13.12, p < 0.005; r2 = 0.37, F1, 20 = 11.99, p < 0.005; r2 = 0.39, F1, 20 = 12.67, p < 0.005, respectively; Fig. 6A). Similarly, CCL2 levels were correlated with time spent in a contracted body position (r2 = 0.58, F1, 17 = 23.57, p < 0.0005; r2 = 0.49, F1, 17 = 16.10, p < 0.001; r2 = 0.54, F1, 18 = 20.88, p < 0.0005, respectively; Fig. 6B). Lastly, CCL2 levels in the frontal cortex, hippocampus, and spleen were each negatively correlated with total distance moved (r2 = 0.29, F1, 19 = 7.80, p < 0.05; r2 = 0.35, F1, 19 = 10.31, p < 0.005; r2 = 0.45, F1, 19 = 15.46, p < 0.001, respectively; Fig. 6C).

Fig. 6

Correlations between LPS-induced CCL2 levels in male C57BL/6J mouse tissues and sickness-like behavior. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. CCL2 was measured via ELISA in brain region and spleen homogenates. Behavioral endpoints that were measured included (A) time spent motionless, (B) time spent in contracted position, and (C) total distance moved. Data are presented as mean ± SEM. Linear regression analysis was used to assess frontal cortex, hippocampus, and spleen CCL2 levels with EPM behavioral endpoints. Linear regression statistics and symbols are provided in figure, and only significant results are shown

CXCL10 levels in the frontal cortex, hippocampus, and spleen were also positively correlated with time spent motionless (r2 = 0.55, F1, 17 = 20.45, p < 0.0005; r2 = 0.47, F1, 21 = 18.25, p < 0.005; r2 = 0.49, F1, 20 = 18.88, p < 0.0005, respectively; Fig. 7A). CXCL10 levels in the frontal cortex, hippocampus, and spleen were also positively correlated with time spent in a contracted body position (r2 = 0.58, F1, 15 = 21.10, p < 0.0005; r2 = 0.70, F1, 19 = 43.55, p < 0.0001; r2 = 0.46, F1, 17 = 14.59, p < 0.005, respectively; Fig. 7B). Finally, CXCL10 levels in the frontal cortex, hippocampus, and spleen were negatively correlated with total distance moved (r2 = 0.56, F1, 16 = 20.37, p < 0.0005; r2 = 0.58, F1, 20 = 27.55, p < 0.0001; r2 = 0.54, F1, 20 = 23.11, p < 0.0005, respectively; Fig. 7C).

Fig. 7

Correlations between LPS-induced CXCL10 levels in male C57BL/6J mouse tissues and sickness-like behavior. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. CXCL10 was measured via ELISA in brain region and spleen homogenates. Behavioral endpoints that were measured included (A) time spent motionless, (B) time spent in contracted position, and (C) total distance moved. Data are presented as mean ± SEM. Linear regression analysis was used to assess frontal cortex, hippocampus, and spleen CXCL10 levels with EPM behavioral endpoints. Linear regression statistics and symbols are provided in figure, and only significant results are shown

Correlation of inflammatory mediators in the brain and spleen with depressive-like behavior

Linear regression analysis revealed significant correlations between chemokine levels in various tissues and depressive-like behavior. Specifically, CCL2 levels in the frontal cortex, hippocampus, and spleen were negatively correlated with latency to immobility in the FST (r2 = 0.36, F1, 14 = 7.80, p < 0.05; r2 = 0.41, F1, 17 = 11.62, p < 0.005; r2 = 0.24, F1, 18 = 5.83, p < 0.05, respectively; Fig. 8A). Additionally, CXCL10 levels in the frontal cortex (r2 = 0.56, F1, 14 = 17.58, p < 0.001) and hippocampus (r2 = 0.41, F1, 19 = 13.01, p < 0.005) were negatively correlated with latency to immobility (Fig. 8B). Notably, CXCL10 levels in the spleen trended towards correlation with latency to immobility (p = 0.051, Fig. 8B).

Fig. 8

Correlations between LPS-induced chemokines in male C57BL/6J tissues and latency to immobility in the FST. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. A CCL2 and (B) CXCL10 were measured via ELISA in brain region and spleen homogenates. Data are presented as mean ± SEM. Linear regression analysis was used to assess frontal cortex, hippocampus, and spleen CCL2 and CXCL10 levels with percentage of time spent immobile and latency to immobility in the FST. Linear regression statistics and symbols are provided in the figure, and only significant results are shown

Effects of β-FNA on NLRP3 in the frontal cortex and spleen

Two-way ANOVA indicated no significant main effect of LPS (F1, 13 = 0.60, p = 0.45) or β-FNA (F1, 13 = 1.49, p = 0.24), as well as no significant interaction of main effects (F1, 13 = 4.24, p = 0.06) on NLRP3 expression in the frontal cortex (Fig. 9A). However, β-FNA + LPS mice tended to have lower levels of NLRP3 than LPS mice.

Fig. 9

Chronic β-FNA effects on LPS-induced NLRP3 expression in male C57BL/6J frontal cortex and spleen tissues. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. RNA was extracted from (A) frontal cortex and (B) spleen tissues, and NLRP3 expression was measured via RT-qPCR. Statistical analysis was performed using the ∆CT-∆CT method and reported as fold change relative to the control group. Data are presented as mean ± SEM. A Two-way ANOVA (n = 3–6/group) indicated no significant main effect for LPS (p = 0.45) or β-FNA (p = 0.24), as well as no significant interaction of main effects (p = 0.06) on NLRP3 expression in the frontal cortex. (B) Two-way ANOVA (n = 4–7/group) revealed a significant main effect for LPS (p < 0.005), no significant main effect for β-FNA (p = 0.50), and no significant interaction of main effects (p = 0.98) on NLRP3 expression in the spleen. Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05)

Two-way ANOVA revealed a significant main effect of LPS (F1, 19 = 12.71, p < 0.005), no significant main effect for β-FNA (F1, 19 = 0.48, p = 0.50), and no significant interaction of main effects (F1, 19 = 0.0004, p = 0.98) on NLRP3 expression in the spleen (Fig. 9B). Pairwise comparisons revealed that expression was similar between saline and β-FNA mice (p = 0.60) and LPS and β-FNA + LPS mice (p = 0.65). However, LPS mice had significantly lower NLRP3 expression in the spleen than saline (p < 0.05) or β-FNA mice (p < 0.01). Additionally, β-FNA + LPS mice had significantly lower NLRP3 expression than β-FNA mice (p < 0.05) and tended to have lower NLRP3 expression than saline mice (p = 0.06).

Effects of β-FNA on IDO1 in the frontal cortex and spleen

Two-way ANOVA revealed a significant interaction of main effects (F1, 13 = 5.81, p < 0.05), but no significant main effect for LPS (F1, 13 = 2.59, p = 0.13) or β-FNA (F1, 13 = 0.25, p = 0.62) on IDO1 expression in the frontal cortex (Fig. 10A). Pairwise comparisons revealed that LPS mice had significantly higher levels of IDO1 expression in the frontal cortex than saline mice (p < 0.05). There was no significant difference between saline and β-FNA mice (p = 0.07), saline and β-FNA + LPS mice (p = 0.17), β-FNA and LPS mice (p = 0.44), β-FNA and β-FNA + LPS mice (p = 0.59), or LPS and β-FNA + LPS mice (p = 0.19).

Fig. 10

Chronic β-FNA effects on LPS-induced IDO1 expression in male C57BL/6J frontal cortex and spleen tissues. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. RNA was extracted from (A) frontal cortex and (B) spleen tissues, and NLRP3 expression was measured via RT-qPCR. Statistical analysis was performed using the ∆CT-∆CT method and reported as fold change relative to the control group. Data are presented as mean ± SEM. A Two-way ANOVA (n = 4–5/group) revealed a significant interaction of main effects (p < 0.05) and no significant main effect for LPS (p = 0.13) or β-FNA (p 0.62) on IDO1 expression in the frontal cortex. B Two-way ANOVA (n = 4–7/group) indicated a significant main effect for LPS (p < 0.01) and β-FNA (p < 0.005), but no significant interaction of main effects (p = 0.87) on IDO1 expression in the spleen. Pairwise comparisons were assessed using Fisher's LSD test; bars with letters in common indicate data are not significantly different (p > 0.05)

Two-way ANOVA indicated a significant main effect for LPS (F1, 18 = 9.14, p < 0.01) and β-FNA (F1, 18 = 10.29, p < 0.005), but no significant interaction of main effects (F1, 18 = 0.03, p = 0.87) on IDO1 expression in the spleen (Fig. 10B). Pairwise comparisons revealed that LPS mice had significantly higher IDO1 expression than saline (p < 0.05) or β-FNA mice (p < 0.0005). While β-FNA + LPS mice tended to have lower IDO1 expression than LPS mice, it fell short of significance (p = 0.06). Pairwise comparisons revealed that β-FNA mice had significantly lower IDO1 expression than saline (p < 0.05) or β-FNA + LPS mice (p < 0.05). IDO1 expression was similar between saline and β-FNA + LPS mice (p = 0.90).

To conclude, while trends showed that β-FNA + LPS mice had lower expressions of IDO1 in the frontal cortex and spleen than LPS mice, it was not significant. Additionally, it appeared that β-FNA treatment suppressed splenic IDO1 expression under control conditions.

Correlation of NLRP3 in the frontal cortex and spleen with behavioral test measures

Linear regression analysis was used to determine whether NLRP3 expression in the frontal cortex and spleen were significantly correlated with anxiety-, sickness-, and depressive-like behavioral measures. Behavioral endpoints analyzed included time in closed arms (EPM), time spent motionless (EPM), time spent in a contracted position (EPM), percentage of time immobile (FST), and latency to immobility (FST). NRLP3 expression in the frontal cortex was not correlated with anxiety-like behavior (F1,10 = 2.64, p = 0.14). However, NLRP3 expression in the frontal cortex increased with sickness-like behavior (as determine by increased time spent motionless and decreased total distance moved; p < 0.05, Fig. 11A). Conversely, linear regression analysis demonstrated that NLRP3 expression in the spleen was negatively correlated with time spent motionless and positively correlated with total distance moved (p < 0.05, Fig. 11B). These findings are consistent with the significantly lower levels of NLRP3 in the spleen of LPS-treated mice compared to the saline-treated mice (Fig. 9B). Overall, NLRP3 expression in the frontal cortex was positively correlated with sickness-like behaviors, while NLRP3 expression in the spleen was inversely related to sickness-like behaviors.

Fig. 11

Correlations between LPS-induced NLRP3 expression and measures of sickness-like behaviors in male C57BL/6J mice. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. NLRP3 was measured via RT-qPCR using frontal cortex and spleen RNA extracts. Linear regression analysis was used to assess (A) frontal cortex and (B) spleen NLRP3 expression with various anxiety-, sickness-, and depressive-like behavioral endpoints. Data are presented as mean ± SEM. Linear regression statistics and symbols are provided in figure, and only significant results are shown

Correlation of IDO1 in the frontal cortex and spleen with behavioral test measures

Linear regression analysis revealed that IDO1 expression in the frontal cortex was significantly correlated with measures of anxiety- and sickness-like behaviors. Specifically, IDO1 in the frontal cortex was correlated with time spent in the closed arms (p < 0.005) and time spent in a contracted position (p < 0.0005, Fig. 12A). Time spent in a contracted position also tended to increase with IDO1 expression in the spleen (p = 0.051, Fig. 12B). While IDO1 levels did not significantly correlate with depressive-like behaviors, latency to immobility in the FST tended to decrease as IDO1 expression increased in the frontal cortex (p = 0.054, Fig. 12C). There were no other significant correlations between IDO1 expression in the frontal cortex or spleen and measures of anxiety-, sickness-, or depressive-like behaviors. Consequently, only IDO1 expression in the frontal cortex significantly correlated with anxiety- and sickness-like behaviors.

Fig. 12

Correlations between LPS-induced IDO1 expression and behavioral measures in male C57BL/6J mice. Micro-osmotic pumps containing saline or β-FNA (42 μg/d) were surgically implanted and dispensed at a flow rate of 0.5 μL/h for 7d. 6d post-surgery, mice (n = 7–8/group) were injected (i.p.) with either 25 μL saline control or LPS (0.83 mg/kg). Behavioral tests were administered 24 h later, and termination followed immediately after. IDO1 was measured via RT-qPCR using frontal cortex and spleen RNA extracts. Linear regression analysis was used to assess frontal cortex and spleen NLRP3 expression with various anxiety-, sickness-, and depressive-like behavioral endpoints. A IDO1 in the frontal cortex correlated with anxiety- and sickness-like behavior. B IDO1 in the spleen trended with sickness-like behavior but fell short of significance. C IDO1 in the frontal cortex trended with depressive-like behavior but fell short of significance. Data are presented as mean ± SEM. Linear regression statistics and symbols are provided in figure