Loss of the interaction between estradiol and insulin-like growth factor I in brain endothelial cells associates to changes in mood homeostasis during peri-menopause in mice

Modulation by exercise of stress resilience in middle age female mice

We first confirmed that our middle-aged (9 months) female population was entering the reproductive senescence transition; over 60% of females already showed altered ovarian cycling -chronic diestrus (Figure 1A). Indeed, a 20% overall reduction in serum E2 as compared to young females, was found at this transitional period (Figure 1A, p<0.001 vs young females).

Middle-aged female mice show markedly reduced immobility time in the forced swim test compared to young females (2 months-old; p<0.001; Figure 1B). Conversely, performance in the elevated plus maze, or in the tail suspension test was similar in young and middle-aged females (Supplementary Figure 1A, B). However, middle-aged female mice did show a detrimental pattern of exercise modulation of these tests. Thus, while in young females exercise did not modify the time spent in the open arms in the elevated plus maze [7], in older females exercise reduced it (p<0.01 vs sedentary; Figure 1C). In addition, while in young females exercise robustly decreased the immobility time in the tail suspension test [7], in middle-aged females exercise did not modify it as compared to sedentary females (Figure 1D).

Since exercise regulation of resilience to stress in young females is accompanied by increased hippocampal IGF-I [7], we next determined whether middle-aged females show changes in hippocampal IGF-I after exercise, and found it to be unaltered (Figure 2A). However, exercise induced an increase in circulating IGF-I in middle-aged females (Figure 2B), while sedentary middle-aged females had a 22% decrease in serum IGF-I levels compared to sedentary younger ones (501 ± 29 ng/ml in young females vs 394 ± 37 ng/ml in middle-aged females; n=6-10, p<0.05).

We determined whether the effects of aging on hippocampal responses to exercise are specific for IGF-I or also affect other proteins reportedly altered in the aging hippocampus during the reproductive senescence transition [9]. We chose two of them, and found that only one, clusterin, but not its receptor Plxna4, showed a distinct age-dependent pattern of responses to exercise (Figure 3).

Estradiol no longer stimulates IGF-I uptake by brain endothelium at middle age

We previously observed that estradiol, acting through its alpha receptor (ERα) on brain vessels, is required to increase hippocampal IGF-I after exercise through increased uptake of IGF-I from the circulation [7]. Indeed, PPT, an ERα agonist, but not DPN, an ERβ agonist, mimicked the stimulatory action of E2 on IGF-I uptake by cultured brain endothelial cells from young female mice (Figure 4A). In contrast, E2 was no longer able to stimulate IGF-I uptake by brain endothelial cells of middle-aged females (Figure 4B). We examined levels of ERα in brain endothelium and found it to be decreased in 9 month old female mice as compared to young ones (Figure 4C). However, expression of IGF-I receptors (IGF-IR) in endothelial cells was unaltered in middle-aged females (Figure 4D). Because ERα and IGF-IR have been shown to interact in response to E2 [10], we next determined whether in middle-aged endothelium this interaction was altered as compared to young endothelium. Indeed, a significantly reduced interaction between these receptors was found (Figure 4E; p<0.05).

Since middle-aged females showed altered ovarian cycling, reduced ERα expression in brain endothelium, and reduced serum E2 levels (see above), we treated them with the ERα agonist PPT during 18 days. We aimed to increase estradiol signaling through this type of receptor because it is through this receptor subtype that E2 interacts with IGF-I. First of all, we observed that submitting middle-aged female mice to the stressful handling produced by daily ip injections of the vehicle was enough to reduce the time spent in the open arm of the elevated plus maze: from around 50 sec in un-injected controls (Figure 1C) to around 10 sec in vehicle-injected females (Figure 5A). However, PPT treatment partially restored the time spent in the open arms (p<0.01 vs sedentary vehicle; Figure 5A), and significantly decreased immobility in the tail suspension test (p<0.05 vs sedentary vehicle; Figure 5B). However, PPT treatment induced a distinct pattern of changes in the performance of behavioral tests after exercise, as compared to young females. Hence, slight, but not significant improvements were seen in the elevated plus maze after exercise, similarly to the effect of the vehicle (Figure 5A). On the other hand, PPT treatment resulted in increased immobility time in the tail suspension test (p<0.05 vs sedentary PPT; Figure 5B). In agreement with a lack of effect of E2 on IGF-I uptake by middle-aged endothelial cells (Figure 4B), all these pattern of changes were independent of hippocampal IGF-I levels, as neither PPT nor exercise affected them (Figure 5C).

Animals and experimental procedures

Adult female C57BL/6J mice (2 and ∼9 months old; Harlan Laboratories, Spain) were housed in standard cages (48 × 26 cm²) with 5 animals per cage. Mice were kept in a room with controlled temperature (22°C) under a 12-12h light-dark cycle and fed with a pellet rodent diet and water ad libitum. All experimental protocols were performed during the light cycle. Animal procedures followed European guidelines (86/609/EEC and 2003/65/EC, European Council Directives) and were approved by the local Bioethics Committee (Government of the Community of Madrid). Thus, following the “reduction” guideline, repeating experiments in young animals was reduced as much as possible as these studies were previously performed by our group (same investigator handling the animals) and have been described elsewhere in detail [7]. We use the data of this recent publication as a reference throughout the manuscript.

Estrous cycles in female mice were monitored by daily inspection of vaginal smears. The material was collected at the same time each day during 10 days. Approximately 10 μl of 0.9% saline were gently flushed into the vagina with the tip of a plastic pipette three times, and the final flush placed onto a glass slide and observed under a light microscope with a 10× objective. Determination of the estrous cycle phase was based on published procedures [30] by examining the proportion of the following cell types: predominance of leukocytes (diestrous), predominance of nucleated epithelial cells (proestrous), predominance of cornified epithelial cells (estrous), and a mix of cell types with a predominance of leukocytes and a few nucleated epithelial and/or cornified squamous epithelial cells (metestrus). The E₂ receptor α agonist (PPT) was administered by intraperitoneal injection at doses of 0.05 gr/kg body weight in corn oil. Treatment started one day before treadmill training and continued throughout the training period. Daily injections were given after running to avoid interference with exercise performance. Controls receive equivalent injections of corn oil.

Experimental design

Middle-aged female mice were randomly divided into the following experimental groups: 1) sedentary control, 2) sedentary control + corn oil, 3) sedentary + PPT (an ERα agonist), 4) exercised control, 5) exercised + corn oil, and 6) exercised + PPT. A subset of these groups (Protocol 1, see Supplemental Figure 1C) underwent several tests to determine resilience to stress and anxiety-like behavior. Specifically, animals were submitted to the forced swim test to determine baseline “depressive-like” behavior. Next, after habituation to the treadmill procedure, mice were submitted to running exercise for two weeks and then to the elevated plus maze followed by a tail suspension test. Another subset (Protocol 2) was submitted to running exercise for 2 weeks and then sacrificed. Their hippocampi and serum were collected to determine gene expression and protein levels. IGF-I levels in hippocampus were determined in all experimental groups whereas clusterin, Plxna4, and serum IGF-I were determined only in groups 1 and 4. In Group 1 we also evaluated serum E2 levels by ELISA to compare them with a young cohort of young female mice.

Treadmill running

Mice were subjected to treadmill running for 2 weeks (5 days/week). Mice were familiarized with the treadmill apparatus (Letica, Italy) to minimize novelty stress and then divided in two groups: exercised and non-exercised. The electrical shock system that encourages the animals to run was disconnected to avoid pain stress. The exercise group ran for 40 min at 12m/min, whereas the control group remained for the same time in the treadmill without running. We chose this mild intensity exercise regime to avoid changes in stress hormones [31] that could interfere with post-exercise behavioral assessment. For biochemical assays, a subset of mice was deeply anesthetized and sacrificed right after the last running session. Trunk blood samples were obtained and brains were perfused with 0.9% saline solution and were either snap frozen for ELISA, Western blot and qPCR or further perfused with 4% paraformaldehyde for immunostaining. Additional groups of animals were used for behavioral testing at different times after running (see Fig 1A).

Elevated plus maze

To assess anxiety-like behavior, animals were placed in the center of an elevated plus maze for 5 min, as described [32]. The maze was at 40 cm from the floor with two opposing protected (closed) arms of 30 cm (length) x 5 cm (wide) x 15.25 (height), and two opposing unprotected (open) arms (30x5x0). Time in the open and closed arms, as well as the number of entries in each arm, were recorded with an automated video-tracking system (Video Tracking Plus Maze Mouse; Med Associates, USA).

Stress resilience

To determine the effects of exercise on resilience to stress, we submitted mice to two different stressful tests, one before and the other after exercise. Thus, three days prior to treadmill training mice were exposed once to forced swimming as a stressor (Supplemental Figure 1C). One day after completion of exercise training, the tail suspension test (TST) was used to measure coping behavior [33]. We used this second type of stressful test to avoid re-exposure of mice to the first stressor (forced swim) and in this way eliminate any potentially confounding learning component. In the forced swim test (FST), mice were placed in a glass cylinder (12 cm diameter, 29 cm height) filled with water (23°C) to a height of 15 cm (to avoid climbing). The duration of the test was 6 min, with the last 4 minutes scored by a blind observer. The animals keep swimming until they give up or they alternated between swimming and floating. In the TST, mice were individually suspended by the tail from a plastic cage (21x26x15) using adhesive tape (distance from tip of tail was 2 cm); the distance from the floor was 35 cm. Animals struggled to get to the floor until they give up and struggled less frequently. A 6 min test session was videotaped and the time spent immobile was scored by a blind observer and referred as percent of total time of duration of the test. In both tests, animals that keep moving for a longer period of time are considered to have a stronger coping strategy

Immunoassays

Western blot (WB) and co-immunoprecipitation (co-IP) assays were performed as described before [34]. Actin was used as a loading control for WB, whereas co-IPs were normalized with the antibody used to pull down. Non-immune serum (NIS) was used as a negative control for IPs while total cells lysate as a positive control. See Supplementary Table 1 for antibodies used. IGF-I in serum and tissues was determined using a species-specific ELISA (R&D Systems, USA), as described in detail elsewhere [35]. E2 was evaluated in serum by ELISA (Enzo Life Sciences, USA), following the manufacturer´s instructions. Blood was collected from the heart after pentobarbital anesthesia before trans-cardiac perfusion.

Endothelial cell cultures

Endothelial cells were obtained as follows: female mice (2 or 9 month-old) were euthanized by C0₂ exposure; the head was removed, immersed in 70% ethanol and transferred to a Petri dish with cool sterile phosphate-buffered saline (PBS; pH 7.4). The brain was removed under sterile conditions and stored on ice with DMEM/F12+10% FBS+penicillin/antimycotic until dissected. Dissection was carried out in cool sterile PBS (pH 7.4) as follows: meninges were removed and discarded; the clean brain tissue was placed on ice in DMEM /F12+10% FBS+ penicillin/antimycotic, cut with a scalpel into small pieces (for no longer than 1 min), and the pieces transferred to a 50 ml Falcon tube with 25 ml of enzymatic medium. The enzymatic medium (previously filtered with a 0.2 μm membrane) consisted of 45000 units of collagenase II (Gibco 17101-015) diluted in 25 ml of PBS + 1% BSA and 25 μl of 1M CaCl2 and 25 μl of 1M MgCl2. Pieces were incubated 2h at 37 ⁰C and then passed through a 20G needle 3-4 times. The resulting mixture was pipetted through a 70 μm disposable cell strainer into a fresh 50 ml tube. Cells were centrifuged at 1300 rpm for 5 min at 4⁰C, the supernatant aspirated and the pellet re-suspended in 25 ml of sterile PBS + 0.1% BSA and again centrifuged at 1300 rpm for 5 min at 4⁰C. The resulting pellet was re-suspended in 1 ml sterile PBS + 0.1% BSA and human endothelial medium plus 20% FBS, 1 μg/ml hydrocortisone, 4 μg/ml puromycin and antibiotic/antimycotic were added. Cells were plated in plates coated with collagen IV/fibronectin and washed with PBS before seeding. The next day, medium was aspirated, cells washed with PBS and the medium replaced with growth medium (DMEM/F12 plus 20% FBS, 1 μg/ml hydrocortisone, 2 ng/ml bFGF, 0.1 mg/ml Cow Brain Extract and penicillin/antimycotic). Medium was then replaced every 2 days.

Biotinylated IGF-I uptake

Endothelial cells obtained from 2 or 9-month old female mice were washed with PBS and serum-free DMEM/F12 without red phenol was added. Three hours later, either 1 nM 17β-estradiol (Sigma, USA), 1 nM PPT (Tocris,USA), or 1 nM DPN, an ERβ agonist (Tocris, USA) were added (or equivalent dose of DMSO in control wells), and 2 hours later biotinylated IGF-I (bIGF-I, IbmH, Germany, 0.2 µg/ml) was added. One hour later, cells were lysed and processed for WB.

qPCR

RNA from endothelial cells was extracted with TRIZOL (Life Technologies, USA) following the manufacturer’s protocol using TaqMan probes. Reactions were performed in an ABI PRISM® 7000 Sequence Detection System. The mRNA levels were normalized with 18sRNA. Primer sequences are shown in Supplementary Table 2.

Statistical analysis

Statistical analysis was performed using GraphPad Prism 5 software (San Diego, CA, USA). All results are shown as mean ± s.e.m. After testing for normal distribution, for single comparisons we used Student’s t-test and for multiple comparisons one or two-way analysis of variance plus Bonferroni’s test. Probability values <0.05 were considered significant.