CGRP overexpression does not alter depression-like behavior in mice

Animals

All animal procedures were performed as previously described (Hashikawa-Hobara et al., 2019) and in accordance with the ARRIVE guidelines and approved by the Animal Care and Use Committee of Okayama University of Science. In keeping with these guidelines, efforts were made to minimize the number of animals used and their suffering. All procedures were approved by the institutional ethics committee of the Okayama University of Science (authorization numbers 2016-03, 2017-08, 2018-04, 2019-08, 2020-08, and 1,421, 1,562). Six-week-old C57BL/6J male mice were purchased from Shimizu Experimental Animals (Shizuoka, Japan) and were habituated to a colony for 2 weeks. CGRP transgenic (Tg) mice were generated using standard techniques, as described previously (Mishima et al., 2018). Eight-week-old C57BL/6J mice and Tg mice were used.

All animals were housed in the Animal Research Center of Okayama University of Science at a controlled ambient temperature of 22 °C, with 50 ± 10% relative humidity, and a 12-h light/dark cycle (lights on at 7:00 AM). A total of 120 mice were used. Eleven mice were used for ELISA assays, 12 mice for mRNA or protein analysis, and 97 mice for the behavioral paradigm. The mice showed variation in behavioral test performance; therefore, the group size was set to approximately 10 mice. Biochemical parameters were usually measured in six animals per group. Animals were group-housed, and each home cage (width 235 mm × depth 353 mm × height 160 mm) contained five to six mice with wood flake litter. When breeding transgenic mice, paper nest boxes were used for environmental enrichment.

Social defeat stress (SDS) paradigm

SDS was performed as previously described (Hashikawa et al., 2017). Briefly, each mouse was attacked by a CD1 aggressor mouse for 10 min a day for 15 consecutive days. After being attacked, each pair of mice was separated by an acrylic board with holes to ensure that visual and olfactory stress would continue to be applied. Although the attacked mice suffered wounds, especially to their backs, we withheld analgesic treatment so as to induce stress. If mice had an abnormal appearance (bleeding, immobility) with little chance of recovery, or rapid and sustained weight loss was observed (20% or more in a few days), the experiment was discontinued and the mice were euthanized. After the stress paradigm, behavioral tests were performed. Mice were then deeply anesthetized and blood samples and hippocampi were collected. A total of 1 h before behavioral tests were performed, the mice were moved to the testing laboratory to acclimatize to the environment.

Behavioral assessments

Analysis of the CGRP levels

Mice were anesthetized with a mixture of three anesthetic agents administered intraperitoneally: medetomidine hydrochloride (Domitol, Meiji Seika Pharma Co., Ltd., Tokyo, Japan, 0.3 mg/kg), midazolam (Dormicum, Astellas Pharma Inc., Tokyo, Japan, 4.0 mg/kg), and butorphanol (Vetorphale, Meiji Seika Pharma Co., Ltd., Tokyo, Japan, 5.0 mg/kg) (Kawai et al., 2011). Blood samples were then collected from the aorta. Serum or hippocampus samples were collected from wild-type or Tg mice, respectively. The CGRP levels were measured using a sandwich enzyme-linked immunosorbent assay (ELISA) kit (Bertin Technologies, Montigny-le-Bretonneux, France).

Quantitative analysis by real-time PCR

The animals were killed by an overdose of pentobarbital-Na (100 mg/kg). Total RNA extraction and real-time PCR were performed as previously described (Hashikawa-Hobara et al., 2021). After extraction, RNA samples were reverse transcribed, real-time PCR was performed using Power SYBR Green PCR Master Mix (Life Technologies Japan Ltd., Tokyo, Japan), and samples were analyzed with the Eco Real-Time PCR System (Illumina Inc., Tokyo, Japan).

Information about the primers, which were designed by the authors, is shown in Table 1. The threshold cycle values for the target (Cgrp, Bdnf, and cFos) and internal control (Actin) genes were determined. Data were collected as previously described (Hashikawa-Hobara et al., 2021). Specifically the fold change of each gene was normalized to Actin and was calculated for each sample relative to the expression in the control samples.

Western blotting

Western blot analyses were performed as previously described (Hashikawa-Hobara et al., 2021). Briefly, mice hippocampus samples were separated by SDS–polyacrylamide gel electrophoresis and then transferred onto a polyvinylidene difluoride membrane (HybondP; GE Healthcare Ltd., Tokyo, Japan). The membrane was incubated with the following primary antibodies: rabbit polyclonal anti-BDNF (1:5,000, Abcam plc, Cambridge, UK.), rabbit anti phosphor-Akt (Ser473) antibody (1:5,000, Cell Signaling Technology, Tokoyo, Japan), rabbit anti Akt antibody (1:5,000, Cell Signaling Technology, Danvers, Massachusetts, USA), rabbit polyclonal anti-mTOR (phosphor Ser2448) (1:5,000, GeneTex, Inc. International Corporation; Funakoshi Co., Ltd., Tokyo, Japan), rabbit polyclonal anti m-TOR (1:5,000, GeneTex, Irvine, CA, USA), rabbit anti phosphor-p70 S6 kinase (Thr 389/412) antibody (1:5,000, Novus Biologicals; Funakoshi Co., Ltd., Tokyo, Japan), and mouse anti p70 S6 kinase antibody (1:5,000, Santa Cruz Biotechnology. Inc., Dallas, TX, USA). After washing with Tris-buffered saline containing 0.1% (v/v) Tween 20, the membranes were incubated with horseradish peroxidase-conjugated secondary antibody (1:20,000) for 1 h at room temperature. The antibody-reactive bands were visualized using a chemiluminescent substrate kit (Immuno star LD, FUJIFILM Wako Chemicals, Tokyo, Japan).

Statistical analysis

GraphPad Prism 9 software (GraphPad Software Inc., San Diego, CA, USA) was used for all statistical analyses. Data analysis of mice was performed blind to group assignment. We used the Grubbs test to remove outliers from groups. All data are expressed as the mean ± standard error of the mean (SEM). Comparisons between two values were analyzed using Welch’s t test. Two-way analysis of variance (ANOVA) was performed when comparing four values. If there was a significant difference in the interaction between groups (stress and CGRP Tg), Tukey’s post-hoc test was used to compare all groups. A p-value < 0.05 was considered significant. In calculating the size effect, Cohen’s d formula was used and 95% confidence interval (CI) values were calculated for the difference between the average values of wild-type and Tg mice.

Characterization of CGRP-overexpressing transgenic mice

We first determined the level of CGRP in CGRP-overexpressing transgenic (Tg) mice. The serum and hippocampal CGRP levels were significantly increased compared with those of wild-type mice (Fig. 1A, serum; p < 0.0001, d = 30.06, 95%CI [1555–1747], hippocampus protein; p = 0.0014, d = 4.09, 95%CI [16.48–40.88], wild-type (n = 6), Tg (n = 5), hippocampus mRNA; p = 0.0045, d = 2.38, 95%CI [425.3–1832], wild-type (n = 6), Tg (n = 6)). Using C57BL6J mice, we previously reported that CGRP intracerebroventricular administration improves depressive-like behavior while increasing the nerve growth factor and c-fos expression (Hashikawa-Hobara et al., 2015). However, whether conditional CGRP overexpression affects the level of neurotrophic factor in mice hippocampi has not been studied. mRNA transcripts for BDNF and c-fos were significantly increased in the hippocampus compared with those of age-matched wild-type mice (Fig. 1B, Bdnf; p = 0.0018, d = 2.28, 95%CI [34.08–126.6], wild-type (n = 6), Tg (n = 6) or Cfos; p = 0.0013, d = 3.38, 95%CI [189.5–530.1], wild-type (n = 5), Tg (n = 5)). Next, we examined the open field test, which is a comprehensive mouse locomotor test. We recorded and analyzed the travel distance (Fig. 1C). Interestingly, locomotor activity of the Tg mice significantly decreased when compared with that of the wild-type (Fig. 1D, p < 0.0001, d = 3.52, 95%CI [−8.068 to −4.595], wild-type (n = 10), Tg (n = 9)). The locomotor activity of one Tg mouse was 14.28, which was an outlier according to the Grubbs test; therefore, it was excluded. We also performed a social interaction test, which assessed whether Tg mice exhibit normal behavior around an unfamiliar mouse. We found that the time spent approaching the aggressor mouse (ICR mouse) was similar to that of the wild-type mice, which indicated that there were no deficits in social interaction (Fig. 1E, p = 0.2936, d = 0.26, 95%CI [−1.048 to 0.614], wild-type (n = 10), Tg (n = 10)).

CGRP-overexpressing mice display stress sensitivity as wild-type mice

Because the expression of BDNF was significantly elevated in the Tg mice hippocampi, we examined whether Tg mice exhibited stress-resistance behavior by performing a battery of behavioral tests. Tg or wild-type mice underwent repeated social defeats during the 15-day social defeat stress procedure. First, we performed the open field test. Mice were randomly divided into four groups: wild-type non-stress (n = 11), wild-type stress (n = 9), Tg non-stress (n = 10), and Tg stress (n = 7). We recorded and analyzed the travel distance or excessive time spent in the center area (Fig. 2A). For distance traveled (locomotor activity), two-way ANOVA showed statistical differences in stress and CGRP Tg interaction (F_{(1, 33)} = 5.890, p = 0.0209) but not stress exposure (F_{(1, 33)} = 2.383, p = 0.1322) and CGRP Tg (F_{(1, 33)} = 3.703, p = 0.0630). Tukey’s post-hoc analysis revealed a significant differences between non-stress wild-type and stress wild-type (p = 0.0285) or non-stress wild-type and non-stress CGRP Tg mice (p = 0.0112). Thus, stress exposure significantly decreased locomotor activity in Tg mice and wild-type mice (Fig. 2B). For time spent in the center area, two-way ANOVA showed statistical differences for stress (F_{(1, 33)} = 10.16, p = 0.0031) and CGRP Tg (F_{(1, 33)} = 5.395, p = 0.0265), but no interaction was detected between stress and CGRP Tg (F_{(1, 33)} = 0.2927, p = 0.5291). Thus, we found that stress and CGRP Tg both shorten the time spent in the center area of the open field test (Fig. 2C). Next, to assess depression-like behavior in Tg mice, we first performed a social interaction test. Although forced swimming or tail suspension tests are commonly used to analyze rodent depressive-like behavior, we decided not to use those experiments because of the significantly reduced locomotor activity of the Tg mice. Therefore, we opted for the widely-used social interaction test and sucrose preference test (Liu et al., 2018) to assess depressive-like behavior. For social interaction rate, two-way ANOVA showed a statistical difference in stress (F_{(1, 31)} = 31.91, p < 0.0001) but not CGRP Tg (F_{(1, 31)} = 1.474, p = 0.2338) and a stress × CGRP Tg interaction (F_{(1, 31)} = 0.8956, p = 0.3513). Thus, we found that the social interaction rate of mice that were exposed to stress was significantly reduced (Figs. 2D, 2E). Next, we performed the sucrose preference test. Mice were randomly divided into four groups with ten mice in each group. Similarly, analysis of anhedonia behavior, which evaluates sucrose preferences, showed statistical differences in two-way ANOVA in stress (F_{(1, 36)} = 23.03, p < 0.0001), but not CGRP Tg (F_{(1, 36)} = 1.137, p = 0.2935) and a stress × CGRP Tg interaction (F_{(1, 36)} = 0.0014, p = 0.994). These results demonstrate that CGRP Tg mice were sensitive to stress.

Stress exposure stimuli inhibited the Akt-mTOR-p70S6K signaling pathway in CGRP-overexpressing mice hippocampi

To assess why CGRP overexpression did not rescue the stress response even though BDNF levels were adequately upregulated, we evaluated hippocampal BDNF expression in both Tg and stressed Tg mice. For BDNF expression, two-way ANOVA showed differences for CGRP Tg (F_{(1, 20)} = 16.96, p = 0.0005) but not for stress (F_{(1, 20)} = 2.702, p = 0.1159) or a stress × CGRP Tg interaction (F_{(1, 20)} = 0.1757, p = 0.6795). We found that stress exposure did not change the levels of BDNF (Fig. 3A). Next, we examined the expression of phosphorylated Akt and mTOR, which form a signaling cascade from BDNF. Interestingly, stress significantly decreased phosphor-Ser473-Akt and phosphor-Ser2448 mTOR in mice hippocampi [Fig. 3B, CGRP Tg (F_{(1, 18)} = 23.17, p = 0.0001), stress (F_{(1, 18)} = 5.601, p = 0.0294), and stress × CGRP Tg interaction (F_{(1, 18)} = 1.707, p = 0.2078), Fig. 3C, CGRP Tg (F_{(1, 18)} = 6.501, p = 0.0201), stress (F_{(1, 18)} = 5.479, p = 0.0310), and stress × CGRP Tg interaction (F_{(1, 18)} = 3.321, p = 0.085)]. To elucidate whether phosphorylated mTOR induces the downstream target p70S6K, we examined phosphorylated p70S6K expression in mice hippocampi. p70S6K is recognized as a kinase for phosphorylation of 40S ribosomal protein S6 and induces increased protein synthesis (Pullen & Thomas, 1997). We observed that stress exposure significantly reduced phosphorylated p70S6K expression [Fig. 3D, CGRP Tg (F_{(1, 19)} = 4.272, p = 0.0526), stress (F_{(1, 19)} = 12.38, p = 0.0023), and stress × CGRP Tg interaction (F_{(1, 19)} = 1.432, p = 0.2461)]. These data demonstrated that CGRP-overexpressing mice are sensitive to stress exposure because phosphorylated Akt/mTOR/p70S6K expression was down-regulated and the subsequent signaling cascades are inactivated even though BDNF levels are abundant.