The LPS produced depression-like, anxiety-like, cognitive impairment and morbid behaviors in mice
To evaluate a possible link between GnRH and inflammation induced depression, we induced acute brain inflammation by injecting bacterial lipopolysaccharides (LPS) or control saline, several behavior tests including depression-like, anxiety-like, cognition behaviors and morbid changes were determined in mice (Fig. 1A). No difference was found in the locomotive activity between the two groups in the open field test (OFT) (t = 1.363, P = 0.1896) (Fig. 1B). Then, depression-like behavior was assessed in mice by using the sucrose preference test (SPT), tail suspension test (TST) and forced swimming test (FST). Compared to the NS mice, the LPS mice showed significantly decreased percent of sucrose preference in the SPT (t = 5.560, P < 0.001) (Fig. 1C), and increased immobility time in the TST (t = 3.576, P = 0.0022) (Fig. 1D) and FST (t = 5.014, P < 0.001) (Fig. 1E) compared to NS mice. Anxiety-like behavior was assessed in mice using the open field test and the elevated plus-maze test. The results showed us that LPS mice spent less time than NS mice in the inner area of the open filed (t = 3.338, P = 0.0037) (Fig. 1F) and in the open arm of the EPM (t = 4.390, P < 0.001) (Fig. 1G). Meanwhile, the novelty object recognition test (NORT) was used to examine the cognitive behavior in mice. Compared to NS mice, LPS mice exhibited significantly decreased discrimination ratio in the NORT (t = 4.380, P < 0.001) (Fig. 1H). At Last, the food consuming and the body weight change were performed to examine the morbid behaviors in mice. LPS mice showed less food consumption (t = 14.90, P < 0.001) (Fig. 1I) and decreased body weight gain (t = 12.63, P < 0.001) (Fig. 1J) compared to NS mice. Taken together, these results indicated that LPS produced depression-like, anxiety-like, cognitive impairment and morbid behaviors in mice.
The alteration of GnRH and GnRHR in the mice after LPS treatment
In order to examine the alteration of GnRH and GnRHR in the mice after LPS treatment, we first examined the expression of GnRH in several depression-related brain regions including hippocampus (HIP), medial prefrontal cortex (mPFC) and hypothalamus (HY) by ELISA. The results showed that the expression of GnRH in the hippocampus of LPS mice was significantly decreased compared with NS group (t = 5.442, P < 0.001) (Fig. 2A), however, there were no significantly difference between NS group and LPS group in medial prefrontal cortex (t = 1.766, P = 0.1154) and hypothalamus (t = 0.6622, P = 0.5228) (Fig. 2B-C). We also used ELISA to detect the amount of serum GnRH, and there were no significantly change between the two groups (t = 1.328, P = 0.2323) (Fig. 2D).
Next, we further examined the alteration of GnRH and GnRHR in hippocampus, medial prefrontal cortex and hypothalamus by Western Blot. The LPS mice exhibited decreased expression of GnRH (t = 2.369, P = 0.0354) and GnRHR (t = 2.256, P = 0.0367) (Fig. 2E-G) in the hippocampus compared to NS mice, while there were no significantly alterations in medial prefrontal cortex (GnRH: t = 1.379, P = 0.1931; GnRHR: t = 0.0585, P = 0.9545 ) (Fig. 2H-J) and hypothalamus (GnRH: t = 1.222, P = 0.3462; GnRHR: t = 1.374, P = 0.1911) (Fig. 2K-M) compared to NS mice. The results of immunohistochemistry showed us that the expression of GnRHR in the hippocampus (CA2, CA3, DG) was decreased compared to NS group (Fig. 2N). All these results indicated that the LPS challenge might result in reduced expression of GnRH and GnRHR in hippocampus specificity.
GnRH agonist triptorelin improved depression-like behavior induced by LPS
Having established that LPS challenges significantly decreased hippocampal GnRH level, we next investigated whether GnRH might affect the depression-like behavior. As we were interested in testing whether GnRH could act peripherally to affect depression-like behavior, we first treated mice with GnRH agonist triptorelin (Trip) via peripheral injections and the experimental flow chart of behavior test was indicating in the Fig. 3A. No differences were found in the locomotive activity among the groups in the OFT (F (2, 27) = 0.2627, P = 0.7709) (Fig. 3B). With regard to the depression-like behavior, one-way ANOVA suggested that there were significantly differences among the groups in the percent of sucrose preference in the SPT [F (2, 27) = 48.81, P < 0.001] (Fig. 3C) and in the immobility time in the TST [F (2, 27) = 9.12, P < 0.001] and FST [F (2, 27) = 52.51, P < 0.001] (Fig. 3D-E). Bonferroni post-test analysis revealed that LPS mice exhibited decreased percent of preference in the SPT compared to NS mice (t = 9.344, P < 0.001), which was reversed by Trip treatment (t = 7.454, P < 0.001). In the FST and TST, Bonferroni post-test analysis revealed that LPS mice showed increased immobility time compared to NS mice (TST: t = 4.014, P < 0.01; FST: t = 7.689, P < 0.001) and Trip treatment could improve this symptom (TST: t = 3.27, P < 0.01; FST: t = 9.712, P < 0.001). With regard to the anxiety behaviors, One-way ANOVA suggested that there were significantly differences among the groups in the time spent in the inner area in the OFT [F (2, 27) = 10.77, P = 0.0004] and in the open arm in the EPM [F (2, 27) = 7.218, P < 0.0031] (Fig. 3F-G). Bonferroni post-test analysis revealed that LPS mice showed decreased time in inner area of the OFT compared to NS mice (t = 4.221, P < 0.001), which was reversed by Trip treatment (t = 3.781, P < 0.01). And Bonferroni post-test analysis revealed that LPS mice exhibited decreased time in the open arm in the EPM (t = 3.366, P < 0.01) compared to NS mice, however, Trip treatment could reverse this change (t = 3.209, P < 0.05). Meanwhile, One-way ANOVA suggested that there were significantly differences in the discrimination ratio among the groups in the NORT [F (2, 27) = 11.42, P = 0.0003] (Fig. 3H). Bonferroni post-test analysis revealed that LPS mice showed significantly decreased discrimination ratio compared to NS mice in the NORT (t = 4.757, P < 0.001), which were improved by Trip treatment (t = 2.775, P < 0.05). One-way ANOVA suggested that there were significantly differences among the groups in food consumption [F (2, 27) = 42.87, P < 0.0001] and body weight gain [F (2, 27) = 32.56, P < 0.0001] during the 24 hour (Fig. 3I-J). Bonferroni post-test analysis revealed that LPS mice consumed less food (t = 8.884, P < 0.001) and showed more serious weight loss (t = 8.026, P < 0.001) during the 24 hours compared to NS mice, and Trip treatment could reverse the weight loss (t = 3.281, P < 0.01) rather than the food consumption (t = 2.183, P > 0.05).
Over-expression of hippocampal GnRH produce anti-depressant effect in LPS treated mice
In order to more specifically assess the relationship between hippocampal GnRH in LPS induced abnormal behavior, the GnRH over-expression lentivirus was stereotactically injected into the ventral hippocampus and the behavior tests were examined 2 weeks later (Fig. 4A). Figure 4B is diagram of viral injection, and we first examined the site of virus injection (Fig. 4C), then the expression of GnRH protein levels in the hippocampus was examined, the result showed us that the GnRH protein levels of LV-GnRH + NS mice was significantly increased compared to LV-GFP + NS mice (t = 7.444, P < 0.01) (Fig. 4D-E). One-way ANOVA suggested that there were significantly differences among the groups in the percent of sucrose preference in the SPT [F (2, 21) = 9.105, P = 0.0014] (Fig. 4F). Bonferroni post-test analysis revealed that LV-GFP + LPS mice exhibited decreased percent of sucrose preference compared to LV-GFP + NS mice in the SPT (t = 3.726, P < 0.01), with the LV-GnRH virus injection, the percent of sucrose preference was increased compared LV-GFP + LPS group (t = 3.665, P < 0.01). One-way ANOVA suggested that there were significantly differences among the groups in the immobility time in the TST [F (2, 22) = 13.78, P = 0.0001] (Fig. 4G) and FST [F (2, 16) = 11.85, P = 0.0007] (Fig. 4H). Bonferroni post-test analysis revealed that LV-GFP + LPS mice showed increased immobility time compared to LV-GFP + NS mice in the TST (t = 2.845, P < 0.05) and FST (t = 4.659, P < 0.001) which was decreased by over-expression of hippocampal GnRH (TST: t = 4.86, P < 0.001; FST: t = 4.613, P < 0.001). One-way ANOVA suggested that there were significantly differences among the groups in the discrimination ratio among the groups in the NORT [F (2, 18) = 17.95, P < 0.0001] (Fig. 4I). Bonferroni post-test analysis revealed that LV-GFP + LPS mice showed significantly decreased discrimination ratio compared to LV-GFP + NS mice in the NORT (t = 4.43, P < 0.001), which were improved by over-expression of hippocampal GnRH (t = 5.589, P < 0.001).
GnRHR inhibitor CEX reversed the protect effect of GnRH agonist in LPS-induced depression model
To confirm that the anti-depression effect of GnRH agonist in LPS treated mice were dependent on its receptor GnRHR, we next determined whether or not selective inhibition of GnRHR via cetrorelix (CEX) would abolish the alleviated depression observed in these mice. Figure 5A is the experimental flow chart of behavior test. No difference was found in the locomotive activity among the groups in the open field test [F (3, 36) = 1.995, P = 0.1321] (Fig. 5B). One-way ANOVA suggested that there were significantly differences among the groups in the percent of sucrose preference in the SPT [F (3, 36) = 31.22, P < 0.0001] and in immobility time in the FST [F (3, 36) = 19.18, P < 0.0001] (Fig. 5C-D). Bonferroni post-test analysis revealed that LPS mice exhibited decreased percent of sucrose preference in the SPT (t = 9.112, P < 0.001) compared to NS mice, which was reversed by Trip treatment (t = 6.116, P < 0.001), however, CEX pretreatment could decrease the percent of sucrose preference (t = 3.261, P < 0.05). Bonferroni post-test analysis revealed that LPS mice showed increased immobility time in the FST compared to NS mice (t = 5.533, P < 0.001) and with the treatment of Trip, the immobility time of LPS + Trip mice was decreased (t = 5.229, P < 0.001) that was reversed by the CEX pretreatment (t = 5.186, P < 0.001). One-way ANOVA suggested that there were significantly differences among the groups in the time spent in the inner area in the OFT [F (3, 36) = 9.73, P < 0.0001] and in open arm of the elevated plus-maze [F (3, 36) = 13.26, P < 0.0001] (Fig. 5E-F). Bonferroni post-test analysis revealed that LPS mice exhibited decreased time in the inner area of the OFT (t = 3.794, P < 0.01) and in the open arm of the EPM (t = 4.565, P < 0.001) compared to NS mice which were all improved by Trip treatment (OFT: t = 2.814, P < 0.05; EPM: t = 5.486, P < 0.001), however, CEX pretreatment could abolish the effect of Trip (OFT: t = 3.631, P < 0.01; EPM: t = 4.010, P < 0.01). These behavioral results from pharmacological and molecular manipulations collectively suggested that GnRH was involved in depression-like behaviors, and GnRH agonist may have the anti-depression effect.
Knockdown hippocampal GnRHR reversed the protect effect of triptorelin in LPS-induced depression model
We next reasoned that if the anti-depression effect of GnRH agonist was indeed associated with hippocampal GnRHR, its depletion should be sufficient to abolish the protective of GnRH agonist. To test this possibility, the GnRHR low-expression lentivirus was stereotactically injected into the ventral hippocampus and the behavior tests were examined 2 weeks later (Fig. 6A). Figure 6B is diagram of viral injection, and we first examined the site of virus injection (Fig. 6C), then the expression of GnRHR protein levels in the hippocampus was examined, the result showed us that the GnRHR protein levels of LV-GnRHR-shRNA + NS mice was significantly decreased compared to LV-GFP + NS mice (t = 3.512, P < 0.05) (Fig. 6D-E). The open field test suggested that there were no significantly differences among the groups in the locomotive activity [F (3, 24) = 1.202, P = 0.3304] (Fig. 6F). One-way ANOVA suggested that there were significantly differences among the groups in the percent of sucrose preference in the SPT [F (3, 36) = 39.26, P < 0.0001] (Fig. 6G) and in the immobility time in the FST [F (3, 31) = 13.84, P < 0.0001] (Fig. 6H). Bonferroni post-test analysis revealed that LV-GFP + LPS mice exhibited decreased percent of sucrose preference compared to LV-GFP + NS mice in the SPT (t = 9.959, P < 0.001) that was enhanced by Trip treatment (t = 8.116, P < 0.001), however, GnRHR knockdown could abolish the effect of Trip treatment (t = 3.687, P < 0.01). Bonferroni post-test analysis revealed that LV-GFP + LPS mice showed increased immobility time compared to LV-GFP + NS mice in the FST (t = 5.521, P < 0.001) which was increased by Trip pretreatment (t = 3.973, P < 0.01), while GnRHR knockdown could reverse the effect of Trip treatment (t = 3.317, P < 0.05). One-way ANOVA suggested that there were significantly differences among the groups in the time spent in the inner area of the open field [F (3, 24) = 12.41, P < 0.0001] (Fig. 5I). Bonferroni post-test analysis revealed that GFP + LPS mice showed decreased time in the inner area of the OFT compared to GFP + NS mice (t = 5.200, P < 0.001) which was enhanced by Trip treatment (t = 4.435, P < 0.01), however, GnRHR knockdown could reverse the effect of Trip treatment (t = 3.148, P < 0.05).
The effect of GnRH agonist on serum testosterone and corticosterone levels in LPS treated mice
To ask if such an anti-depression effect of GnRH agonist is due to its peripheral endocrine effects, we first examined whether the serum testosterone and corticosterone levels of mice would be altered after GnRH agonist treatment in LPS mice. The result showed us that there were no significantly differences among the groups in the amount of testosterone [F (2, 18) = 0.3182, P = 0.7315] (Fig. 7A) and corticosterone [F (2, 18) = 2.096, P = 0.1520] (Fig. 7B). Considering that the triptorelin may affect the reproductive system, we also examined the total sperm count and sperm motility in the mice, the results showed us that there no significantly differences in the total sperm count [F (2, 17) = 1.511, P = 0.2488] (Fig. 7C) and sperm motility [F (2, 17) = 1.116, P = 0.3505] (Fig. 7D) among the three groups. Although these data do not exclude the possibility that GnRH agonist mediated HPG function might contribute to its function, they propose that GnRH agonist functions in the local brain environment. Collectively, these results suggest that the influence of the GnRH on neuro-immune system might be related to its neuronal effect rather than its endocrine influence.
The effect of GnRH agonist on serum and hippocampus inflammation levels in LPS-injected mice
Then we examined the expression of inflammatory cytokines in serum and hippocampus of mice. One-way ANOVA suggested that there were significantly differences among the groups in several inflammatory cytokines in the serum including IL-1β [F (2, 27) = 7.417, P = 0.0027], IL-6 [F (2, 24) = 8.729, P = 0.0014]) and TNF-α [F (2, 35) = 23.22, P < 0.0001] (Fig. 8A). Bonferroni post-test analysis revealed that LPS mice exhibited significantly increased expression IL-1β (t = 3.175, P < 0.05), IL-6 (t = 3.877, P < 0.01) and TNF-α (t = 4.952, P < 0.001) compared to NS mice, which were effectively reversed by Trip pretreatment (IL-1β: t = 3.476, P < 0.01; IL-6: t = 2.926, P < 0.05; TNF-α: t = 6.507, P < 0.001). With regard to expression of inflammatory cytokines in hippocampus, One-way ANOVA suggested that there were significantly differences among the groups in the expression of IL-1α mRNA [F (2, 27) = 23.19, P < 0.0001], IL-1β mRNA [F (2, 27) = 14.42, P < 0.0001], IL-6 mRNA [F (2, 27) = 7.12, P = 0.0033] and TNF-α mRNA [F (2, 27) = 38.33, P < 0.0001] (Fig. 8B). Bonferroni post-test analysis revealed that LPS mice exhibited significantly increased expression of IL-1α mRNA (t = 6.806, P < 0.001), IL-1β mRNA (t = 5.272, P < 0.001), IL-6 mRNA (t = 3.237, P < 0.01) and TNF-α mRNA (t = 8.191, P < 0.001) compared to NS mice and all these changes were reversed by Trip pretreatment (IL-1α mRNA: t = 3.607, P < 0.01; IL-1β mRNA: t = 3.524, P < 0.01; IL-6 mRNA: t = 3.299, P < 0.01; TNF-α: mRNA: t = 7.774, P < 0.001).
We further examined the expression of inflammatory cytokines in hippocampus by Western blot. One-way ANOVA suggested that there were significantly differences among the groups in the expression of IL-1α [F (2, 13) = 12.20, P = 0.0010], IL-1β [F (2, 12) = 10.59, P = 0.0022] and IL-6 [F (2, 10) = 15.73, P = 0.0008] in the hippocampus (Fig. 8C-D). Bonferroni post-test analysis revealed that LPS mice exhibited significantly increased expression IL-1α (t = 3.663, P < 0.01), IL-1β (t = 4.389, P < 0.01) and IL-6 (t = 3.084, P < 0.05) compared to NS mice, which were effectively reversed by Trip pretreatment (IL-1α: t = 4.603, P < 0.01; IL-1β: t = 3.396, P < 0.05; IL-6: t = 5.607, P < 0.001).
Effect of triptorelin on the activation of microglia in LPS-injected mouse hippocampus
Microglia are the brain's innate immune cells, monitor and protect the brain, so we also examined the alteration of microglia marker IBA-1 in the hippocampus by immunohistochemistry. One-way ANOVA suggested that there were significantly differences among the groups in the number of IBA-1 positive cells [F (2, 24) = 35.93, P < 0.0001] in the hippocampus (Fig. 9A-B). Bonferroni post-test analysis revealed that LPS mice exhibited significantly increased number of IBA-1 positive cells compared to NS mice (t = 8.373, P < 0.001), with the treatment of Trip, IBA-1 positive cells was decreased (t = 6.061, P < 0.001). Next, we determined the morphological alterations of microglia (Fig. 9C). One-way ANOVA suggested that there were some differences among the groups in intersection number of microglia arborization (F (7, 14) = 49.38, P < 0.0001) in representative skeleton analysis (Fig. 9D). Bonferroni post-test analysis revealed that intersection number of microglia arborization significantly decreased in the LPS mice compare to Vehicle group (12: t = 2.81, P < 0.01; 15: t = 3.2, P < 0.01; 18: t = 2.71, P < 0.01; 21: t = 3.24, P < 0.01), which was increased after Trip treatment (12: t = 2.07, P < 0.05; 15: t = 2.75, P < 0.01; 18: t = 2.26, P < 0.05; 21: t = 1.61, P > 0.05). We also examined microglia markers CD68 and CD11b by Western Blot, One-way ANOVA suggested that there were significantly differences among the groups in the expression of CD68 [F (2, 13) = 13.56, P = 0.0007] and CD11b [F (2, 24) = 5.222, P = 0.0131] in the hippocampus (Fig. 9E-F). Bonferroni post-test analysis revealed that LPS mice exhibited significantly increased expression of CD68 (t = 2.821, P < 0.05) and CD11b (t = 2.784, P < 0.05) compared to NS mice which were reversed by Trip treatment (CD68: t = 5.181, P < 0.001; CD11b: t = 2.813, P < 0.05).
Once activated by neuroinflammation, microglia shows morphology change and producing both neurotoxic molecules and neurotrophic factors depending on their pro-inflammatory M1 and anti-inflammatory M2 phenotypes. Thus, we further determined the expression of M1 (CD16, CD86, CXCL10) and M2 markers (Arg1, CD206). One-way ANOVA suggested that there were significantly differences among the groups in the expression of CD16 mRNA [F (2, 22) = 8.442, P = 0.0019], CD86 mRNA [F (2, 23) = 7.120, P = 0.0039], CXCL10 mRNA [F (2, 22) = 17.73, P < 0.0001] and CD206 mRNA [F (2, 23) = 9.057, P = 0.0013] rather Arg1 mRNA [F (2, 23) = 1.909, P = 0.1710] in the hippocampus (Fig. 9G-H). Bonferroni post-test analysis revealed that LPS mice exhibited significantly increased expression of CD16 mRNA (t = 4.102, P < 0.01), CD86 mRNA (t = 3.713, P < 0.01), CXCL10 mRNA (t = 5.395, P < 0.001) and CD206 mRNA (t = 4.057, P < 0.01) compared to NS mice, the Trip treatment could decreased the expression of CD16 mRNA (t = 2.656, P < 0.05), iNOS mRNA (t = 1.588, P < 0.05), CD86 mRNA (t = 2.820, P < 0.05) and CD206 mRNA (t = 3.516, P < 0.01) rather CXCL10 mRNA (t = 1.417, P > 0.05).
Triptorelin reversed LPS-induced hippocampal neurogenesis suppression
Over-activated microglia is associated with deficits in neurogenesis that contributes to the physiopathology of major depressive disorder. Thus, we examined the alteration of adult neurogenesis marker DCX in the hippocampus. One-way ANOVA suggested that there were significantly differences in the number of DCX-positive cells among the groups [F (2, 36) = 5.912, P = 0.0060] (Fig. 10A-D). Bonferroni’s post hoc test revealed that the number of DCX-positive cells in LPS group were significantly decreased compared to NS group (t = 4.641, P < 0.01), which was reversed by Trip treatment (t = 3.680, P < 0.05).