Insulin and insulin-like growth factor-I receptors in astrocytes exert different effects on behavior and Alzheimer´s-like pathology

Animals

Mice were used according to Animal Research: Reporting of in vivo Experiments (ARRIVE) and this study is reported in line with the guidelines (Zegarra-Valdivia et al., 2022). Transgenic mice with tamoxifen-regulated deletion of IGF-IR or IR in astrocytes (GFAP-IGF-IR KO and GFAP-IR KO mice, respectively) were obtained as described (Garcia-Caceres et al., 2016; Noriega-Prieto et al., 2021) crossing IR^f/f (B6.129S4(FVB)-Insr^tm1Khn/J RRID:IMSR, Jackson labs; stock number 006955) or IGF-IR^f/f (B6, 129 background; Jackson Labs; stock number: 012251) with hGFAP-CreER^T2 mice (C57B&/6xSJL/J mix background Jackson Labs, stock number: 012849). To knock down the target gene, tamoxifen was administered to 2- months old mice for 5 days (75 mg/kg, Sigma, intraperitoneally) as described (Hirrlinger et al., 2006), and animals were used one month later. Controls littermates received the vehicle (corn oil). GFAP-IGF-IR KO and GFAP-IR KO display reduced mRNA levels in brain, as reported by Noriega-Prieto et al. (2021) and Garcia-Caceres et al. (2016). GFAP-IR KO mice show brain IGF-IR levels similar to wild type mice whereas GFAP-IGF-IR KO mice had normal brain IR levels (Hernandez-Garzon et al., 2016). APP_swe and PS1Δ9 mice of C57BL6/J background were from the colony of the Cajal Institute. Homozygous APP/PS1 mice were crossed with homozygous GFAP IGF-IR KO or GFAP IR KO mice to obtain the respective compound strains. Studies were carried out at the age of 10-11 months-old, when pathology is well developed.

Ethical considerations

Mice were were housed in standard cages (48 × 26 cm²) with 5 mice per cage. Mice were maintained on a light-dark cycle (12-12 h, lights on at 8 am) at constant temperature (22°C) and humidity, and with food (pellet rodent diet) and water ad libitum. All experimental protocols were performed during the light cycle and followed European guidelines (86/609/EEC & 2003/65/EC, European Council Directives).

Studies were approved by the respective local Bioethics Committees (Government of the Community of Madrid, MERGEFIELD CÓDIGO PROEX 193.4/20 (2020) and UPV M20_2021_168 (2021). Animals were not randomized and were used in a sex-balanced manner throughout. Potential confounders were not accounted for. Each experimenter took account of group allocation under study. All efforts were done to reduce harm to the animals. Mice were handled for three days prior to any experimental manipulations and familiarized with behavioral arenas to minimize novelty stress or deeply anesthesized with pentobarbital prior to sacrifice, when needed. Sample sizes were kept as little as possible to comply with current animal reduction policies. No adverse events were expected, nor found. End-point measures included checking reflexes in deeply anesthesized animals prior to culling.

Behavioral tests

These tests were used to determine behavior under laboratory-controlled conditions. These are observational studies with no a priori hypothesis.

Barnes maze. To assess spatial learning and memory, animals received reinforcement to escape from an open circular platform (92 cm Ø with 20 holes of 5 cm Ø) to the “escape chamber”, as described (Ortiz et al., 2010; Zegarra-Valdivia et al., 2019). All animals received appropriate training (four trials per day), and trials were separated by 15 min. After each trial, the maze was cleaned with 70% alcohol. On the 5^th day, both groups were tested, and once more 48 hours later, evaluating the long-term memory of the animals. Time to escape to the safe chamber was quantified.

Open field. Exploratory behavior and locomotion were assessed by introducing the animal to an open field arena (42 cm × 42 cm × 30 cm, Versamax; AccuScan Instruments, Inc.) for 10 min. All parameters were quantified as described (Zegarra-Valdivia et al., 2019). Time spent exploring specific areas of the arena was measured.

Elevated plus maze. To assess anxiety-like/coping behavior, mice were introduced in a maze of 40 cm from the floor with two opposing arms. Two protected (closed) arms (30 cm (length) × 5 cm (wide) × 15.25 (height), and two opposing unprotected (open) arms (30 cm (length) × 5 cm (wide). Each animal was introduced in the middle of the apparatus for 5 minutes. Stress was scored as time spent in the closed arms while coping behavior was estimated by time spent in the open arms. All measures were recorded (Video Tracking Plus Maze Mouse; Med Associates, USA), and analyzed as described (Munive et al., 2019).

Y-maze. This test measures spontaneous alternation as an index of working memory (Sarter et al., 1988). The maze is made of black-painted wood, and each arm is 25 cm long, 14 cm high, 5 cm wide, and positioned at equal angles. The mouse is placed at the end of one arm to move freely from side to side of the maze during an 8-min session. Videos recorded the sequence of entries during the whole time of the experiment and were analysed off-line. Entrance to each arm is scored when the mouse places the hind paws entirely in the zone. Alternation was defined as successive entries into the three arms on overlapping triplet sets. Consecutive triplets were analyzed, and alternate behaviour was calculated as the percentage of actual alternation (number of triplets with non-repeated entries) versus total alternation opportunities (total number of triplets), as described (Recinto et al., 2012; Yan et al., 2017).

Tail suspension. In this test coping behaviors are determined. As already described (Munive et al., 2019), mice were suspended by the tail from a plastic cage (21×26×15) with 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 time spent immobile was scored and referred as percent of total time of duration of the test.

Forced swim. This test measures depressive-like behavior. As described (Munive et al., 2019), 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) and videotaped. The test lasted 6 min, and immobility time was scored the last 4 minutes.

Spatial Y-maze. This test was used for spatial, novelty-seeking, and short-term memory assessment by measuring time spent in the novel arm (Hausrat et al., 2015; Biundo et al., 2018). As before, the maze was made of black-painted wood and each arm was 25 cm long, 14 cm high, 5 cm wide and positioned at equal angles. Each mouse was allowed to explore two arms of the Y-maze apparatus during the first trial (training) for 5 min. One hour later, the third arm was opened, and the mouse was returned to the same maze and allowed to explore all the three arms (testing). Visual cues were used to guide environment exploration, as described (Biundo et al., 2018). Animals with preserved cognition remember the previously blocked arm and they will enter it first on the second trial and spend more time exploring it. Distribution of mice and novel arms were balanced within each group. We cleaned the maze with 70% ethanol to remove olfactory cues between trials.

Rota-rod. Motor coordination was assessed with the rota-rod test, as described before (Fernandez et al., 1998). Briefly, mice were submitted to 1 min training session in the immobile apparatus. When the mouse falls, it is placed back into the rotating rod. Thereafter, mouse performance was tested in 5 min sessions every 15 min in 4 trials with increasing acceleration up to 40 rpm. The latency to fall off the rod in the final trial was measured and compared between groups.

Social behavior. Social behavior includes rewarding and motivational processes (Trezza et al., 2011; McCall and Singer, 2012). We studied social affiliation and social novelty/preference as described by others (Kaidanovich-Beilin et al., 2011). We placed each mouse in a cage with three compartments (one central and two lateral arms); in each compartment, we added a grid with one stranger mouse or an empty grid to assess social affiliation (intention to stay with the same species). We leave the mouse to explore for 10 minutes and record the time of direct interaction. Then, we cleaned the three chambers with ethanol (70%) to eliminate olfactory cues and placed the mice again in the center chamber. We include the previous stranger mice in the same arm (now named “familiar mouse”). In the empty space we include a new mouse (“stranger mouse”) and leave the animal free to explore and record the time of direct interaction.

Immunocytochemistry

Immunocytochemistry was performed as described in detail before (Fernandez et al., 2012). A pre-treatment of 70% formic acid was used before incubation with anti-human Aβ antibody (1:50, Dako clone 6F/3D). Primary antibody was omitted as control. Confocal analysis was performed in a Leica (SP5 Direct, Germany) microscope. For plaque morphometry, 1-4 vibratome brain sections (50 μm, parietal cortex and hippocampus) were used to assess the density of Aβ plaques using Imaris software (Vs 9.0.2) (RRID:SCR_007370). Measurements were done as explained in detail elsewhere (Fernandez et al., 2012). Briefly, images were recorded using a 5X objective and were converted to gray scale to improve the contrast between signal and noise. All pictures were measured separately applying the same threshold. Areas were measured inside a reference circle in the hippocampus or parietal cortex with a standard size of 300 mm². We then calculated the percentage of reference area occupied by Aβ plaques.

Statistics

The number of animals for each experiment was calculated according to past experience with no hypothesis-driven outcomes, as these are observational studies. All animals in each group were included in analyses with no exclusion criteria applied a priori. Values were relativized compared to the control or baseline condition. Results are expressed as the average of the relative values obtained in each independent test (mean ± standard error) for each experiment and analyzed with GraphPad Prism 8.0 software (RRID:SCR_002798) (alternative open access program: R Program). Normality was confirmed using the Shapiro-Wilk normality test and equal variances with Levene’s test. Later, student’s t-test was used for comparison of two groups, or ANOVA for comparison of more than two groups with a Tukey or Sidak’s post-hoc analysis. Further details are explained in each figure. A statistically significant difference was considered when p<0.05.

Behavioral traits in mice lacking insulin or IGF-I receptors in astrocytes

Recent publications in different models of downregulation of either insulin or IGF-I receptors in astrocytes have started to unveil specific actions of these receptors in this type of glial cells (Cai et al., 2018; Logan et al., 2018; Manaserh et al., 2019; Noriega-Prieto et al., 2021). We confirm that adult GFAP IR KO mice gradually show a depressive-like phenotype (Cai et al., 2018), as determined by the forced swim and the tail suspension tests. These alterations are seen in adult (>6 months old), but not younger mice (Figure 1A-B) (Zegarra-Valdivia et al., 2022). The existence of a depressive-like phenotype was reinforced by the observation that adult GFAP IR KO mice show disturbed responses to social novelty (Figure 1C), although not to social affiliation (Figure 1D). As determined in the open field test and elevated plus maze, GFAP IR KO mice did not show changes in anxiety levels either, which are frequently associated to depression (Figure 1E-F). These mice have intact cognition, as determined in the Barnes and Y maze tests assessing learning and memory (Figure 2A-C). In addition, GFAP IR KO mice did not show deficits in ambulation or motor coordination (Figure 2C-D).

Figure 1. Mood traits in mice lacking insulin receptors (IR) in GFAP astrocytes (GFAP-IR KO mice).

A, Adult (right histograms, control n= 6, GFAP IR KO n=7, t-test, t= 2.54; *p<0.05), but not young (left histograms, control n= 8, GFAP IR KO n=7, t-test, t= 1.26, p=0.22) GFAP IR KO mice show increased immobilization time in the tail suspension test (upper drawing), an indicator of a depressive-like behavior and reduced resilience to stress. B, Similarly, in the forced swim test (upper drawing), adult (right histograms, control n= 5, GFAP IR KO n=6, t-test, t= 2.5; *p<0.05, Welch`s correction), but not young (left histograms, control n= 5, GFAP IR KO n=6, t-test, t= 0.10; p=0.922) GFAP IR KO mice show increased depressive-like performance, with less time spent swimming. C, Social novelty, as measured by time spent with a novel partner vs a familiar one (upper drawing), was impaired in GFAP IR KO mice (control n= 10, GFAP IR KO n=12, t-test, t= 2.25; *p<0.05). D, Social affiliation, as determined by time spent with a stranger mouse vs an empty cage, was normal in GFAP IR KO mice (control n= 10, GFAP IR KO n=12, 2-way RM ANOVA, condition factor, F(1,20)=28.74; ***p<0.001, Sidak's multiple comparisons test, control familiar mice vs empty cage, ***p<0.001, GFAP IR KO novel vs familiar, **p<0.01). E, Time spent in the center of an open arena (upper drawing), a measure of novelty stress indicating levels of anxiety remained within control levels in adult GFAP IR KO mice (n=12 per group; t-test; t=0.77, p=0.445). F, Anxiety levels, as determined by time in the open arms of the elevated plus maze (upper drawing), are slightly, were normal in adult GFAP IR KO mice (control n= 21, GFAP IR KO n=24, t-test, t= 1.46, p=0.15). GFAP=glial fibrillary astrocytic protein.

Figure 2. Cognition in GFAP-IR KO mice.

A, Adult GFAP IR KO mice performed similarly in the Barnes maze as compared to littermates, indicating intact spatial learning (n=9 per group, training days: 2-way ANOVA: F(1,36)=0.10, p=0.74; test day: t-test, Welch’s correction, t=0.65, p=0.53). B, Time spent in the novel arm of the spatial Y maze was similar to littermates in GFAP IR KO mice (n=9 per group, t-test, t=1.24, p=0.23). C, Number of spontaneous alternations in the arms of the Spontaneous Alternation Y maze, a measure was similarly unaltered in adult GFAP IR KO mice (n=9 per group, t-test, Welch’s correction, t= 0.67, p=0.51). D, No differences were observed in horizontal (left histograms) and vertical (right) activity in the open field arena was observed between experimental groups (n=12 per group, H: t-test, t=0.84, p=0.40; V: t-test, t= 0.05, p=0.95). E, Control littermates and GFAP IR KO mice show similar levels of motor coordination, as assessed in the rota-rod (control n=12, GFAP IR KO n=14, H: t-test, t=1.52, p=0.14). IR=insulin receptors, GFAP=glial fibrillary astrocytic protein.

Conversely, adult (>6 months old) GFAP IGF-IR KO mice show specific impairments in spatial memory as assessed in the Barnes and Y mazes (Figure 3A-B), confirming previously observed deficits in cognition in these mice (Noriega-Prieto et al., 2021). However, working memory, as assessed by the alternation ratio in the Y maze, was intact (Figure 3C). These mice show normal social affiliation, whereas their preference for a novel partner was slightly impaired (Figure 3D-E). GFAP IGF-IR KO mice did not show mood disturbances either, as determined by time spent in the center of an open arena or in the open arms of the elevated plus maze (Figure 4A-B). GFAP IGF-IR KO show normal ambulatory behavior in the open field (Figure 4C), and in motor coordination tested in the rota-rod (Figure 4D).

Figure 3. Cognitive traits in mice lacking IGF-I receptors (IGF-IR) in GFAP astrocytes (GFAP-IGF-IR KO mice).

A, Spatial learning in the Barnes maze (upper drawing) was markedly affected in GFAP IGF-IR KO mice, showing significantly reduced memory (control n= 12, GFAP IR KO n=19, training days: 2-way ANOVA, time factor, F(3,122)=12.7; ***p<0.001, Sidak's multiple comparisons test, control vs GFAP 4^th day of training, *p<0.05; test day: Mann-Whitney U: 21.5, ***p<0.001). B, Time spent in the novel arm of the Y maze (upper drawing), a measure of spatial memory, was reduced in GFAP IGF-IR KO mice (control n= 8, GFAP IGF-IR KO n=12, t-test, t= 2.26, *p<0.05). C, Number of spontaneous alternations in a Y maze, a measure of working memory (upper drawing), was similarly unaltered in adult GFAP IGF-IR KO mice (n=8 per group; t-test; t=0.98, p=0.342). D, Social affiliation, as determined by time spent with a stranger mouse vs an empty cage (upper drawing), was normal in GFAP IGF-IR KO mice (control n=9, GFAP IGF-IR KO n=8, 2-way RM ANOVA, condition factor, F(1,15)=19.13; ***p<0.001, Sidak's multiple comparisons test, control familiar mice vs empty cage, **p<0.01, GFAP IR KO familiar vs empty cage, *p<0.05). E, Social novelty, as measured by time spent with a novel partner (upper drawing), was impaired in GFAP IGF-IR KO mice (control n=9, GFAP IGF-IR KO n=8, 2-way RM ANOVA, condition factor, F(1,15)=11.18; **p<0.01, Sidak's multiple comparisons test, control novel mice vs familiar mice, *p<0.05, GFAP IGF-IR KO novel mice vs familiar mice, p=0.16). GFAP=glial fibrillary astrocytic protein.

Figure 4. Mood homeostasis in GFAP IGF-IR KO mice.

A, No differences in time spent in the center of an open field were observed between littermates and mutant GFAP IGF-IR KO mice (control n=28, GFAP IGF-IR KO n=22, t-test, t=0.73, p=0.46). B, No differences were observed in anxiety levels determined in the EPM between GFAP-IGF-IR KO and littermates (control n=16, GFAP IGF-IR KO n=15, t-test, t=0.48, p=0.63). C, No differences were observed in horizontal (left histograms) and vertical (right) activity in the open field arena was observed between GFAP IGF-IR KO mice and littermates (control n=28, GFAP IR KO n=22, H: t-test, t=0.61, p=0.54; V: t-test, t=0.12, p=0.90). D, Control and GFAP-IGF-IR KO mice show similar levels of motor coordination, as assessed in the rota-rod (control n=22, GFAP IR KO n=27, t-test, t=0.22, p=0.82).

Modulation of Alzheimer’s-like pathology in mice lacking insulin or IGF-I receptors in astrocytes

Mice lacking IGF-IR in neurons show reduced AD-like pathology when cross-bred with a mouse AD model (Gontier et al., 2015), whereas mice lacking IR in neurons have not shown changes in AD-like pathology (Freude et al., 2009b). To analyze possible cell-dependent actions of these receptors in AD-like pathology, we crossed either GFAP IR KO or GFAP IGF-IR KO mice with APP/PS1 mice to obtain compound mutants and determined the impact of these receptors in memory loss associated to AD pathology seen in this mouse model. We observed that double GFAP IR KO/APP-PS1 mice presented significantly greater working memory loss compared to controls, as indicated by reduced spontaneous alternation in the Y maze (Figure 5A). In contrast, double GFAP IGF-IR KO/APP-PS1 showed enhanced cognition when compared to APP/PS1 mice, but no changes when compared to vehicle-treated GFAP IGF-IR/APP-PS1 mice (Figure 5B). Importantly, vehicle-treated control with preserved IGF-IR activity in astrocytes also showed enhanced cognition when compared to APP/PS1 mice (Figure 5B). Associated to greater memory loss we observed greater amyloid load in GFAP IR KO/APP-PS1 mice (Figure 5C), while in GFAP IGF-IR KO/APP-PS1 mice changes in amyloid plaque load were, again, control-dependent (Figure 5D). When compared to APP/PS1 mice, no changes were seen, but when compared to vehicle-treated controls, plaque load was increased. Of note, vehicle-treated controls show reduced plaque load when compared to APP/PS1 controls (Figure 5D).

Figure 5. Modulation of Alzheimer´s-like pathology in GFAP IR KO/APP-PS1 and GFAP IGF-IR KO/APP-PS1 mice.

A, Performance in the working memory version of the Y maze was impaired in GFAP IR KO APP/PS1 mice (n=8 per group; One-way ANOVA, F=73.23; ***p<0.001; Tukey’s Multiple comparison test, APP/PS1 vs. GFP IR KO-APP/PS1: ***p<0.001, GFP IR Control-APP/PS1 vs. GFP IR KO-APP/PS1: ***p<0.001). B, Working memory determined in the Y maze remained unaltered in GFAP IGF-IR KO APP/PS1 mice and controls (n=8 per group; One-way ANOVA, F=2.9; p=0.07). C, Amyloid plaques in the parietal cortex and hippocampus in GFAP IR KO/APP-PS1 mice and controls. Representative photomicrographs showing amyloid plaques (red). Histograms show number of plaques/μm² in three experimental groups (n=8 per group; One-way ANOVA, F=25.78; ***p<0.001; Tukey’s Multiple comparison test, APP/PS1 vs. GFP IR KO-APP/PS1: ***p<0.01, GFP IR Control-APP/PS1 vs. GFP IR KO-APP/PS1: ***p<0.001). D, Amyloid plaques in the parietal cortex and hippocampus in GFAP IGF-IR KO/APP-PS1 mice and controls does not show differences between groups. Representative photomicrographs showing amyloid plaques (red). Histograms show number of plaques/μm² in the three experimental groups (n=8 per group; One-way ANOVA, F=1.35; p=0.32). IR=insulin receptors, GFAP=glial fibrillary astrocytic protein.

Underlying data

Harvard Dataverse: DATA SET - ASTROCYTE INSULIN AND INSULIN-LIKE GROWTH FACTOR I (IGF-I) RECEPTORS. https://doi.org/10.7910/DVN/Y7K97E (Zegarra-Valdivia et al., 2022).

This project contains the following underlying data: