Humans
Frozen tissues of the prefrontal cortex [Broadman areas 9–10] (n = 79) subjects with age at death ranging from 29 to 98 years (yr), and postmortem interval (PMI) ≤ 35 h (h) were obtained from the following brain banks: Rush Religious Orders Study (RROS), Center for Neurodegenerative Disease Research, University of Pennsylvania School of Medicine, Harvard Brain Tissue Resource Center and the Emory Center for Neurodegenerative Disease, Emory University School of Medicine. A total of n = 21 control subjects (12 M/9F) were clinically examined and diagnosed with no cognitive impairment or insufficient to meet criteria for dementia (age: 69.8 ± 3.6 yr, PMI: 13 ± 1.8 h, Braak stage: 1.3 ± 0.23). Neurodegenerative disorder cases include n = 40 AD (14 M/26F, age: 78.6 ± 1.8 yr, PMI: 11.3 ± 0.9 h, Braak stage: 5.6 ± 0.07) and n = 18 Parkinson’s disease (PD; 12 M/6F, age: 75.4 ± 2.13 yr, PMI: 8.9 ± 1 h, Braak stage: 1 ± 0.17). Cognition was assessed within the year prior death using the Mini-Mental State Exam (MMSE). Scores are 29.2 ± 0.35 for controls, 12.2 ± 1.3 for AD and 26 ± 1 for PD. Exclusion criteria for AD cases included argyrophilic grain disease, frontotemporal dementia, Lewy body disease, mixed dementias, PD, and stroke. Neuropathology was determined by a board certified neuropathologist blinded to the clinical diagnosis. Tissue samples were processed as previously described [22] for Western blotting using the following antibodies: p-GR[S267], p-GR[S134], p-GR[S211], p-GR[S226] (1:1000, all made by M. Garabedian, New York University Grossman School of Medicine (NYUGSOM; NY, NY, USA [7]), total GR (1:400, P20, Santa Cruz Biotechnology, Santa Cruz, CA, USA), APP (1 μg/ml, PA1-84,165) and anti-oligomer A11 (1 μg/ml, ThermoFisher, Waltham, MA, USA), Tau-1 (1:1000, Sigma-Aldrich MAB3420, USA), PHF1[p-Tau S396/S404] (1:500, a gift of P. Davies, Long Island Jewish Medical Center, Northwell Health, New Hyde Park, NY, USA), GAPDH (1:1000, Meridian Life Sciences, Menphis, TE, USA), FKBP5 (1:1000, Abcam ab2901, Paris, France), TrkB (1:1000, 610,101, BD Biosciences, USA), p-TrkB[Y816] (1:1000, a gift of M.V. Chao, NYUGSOM), BDNF (1:400, N20, Santa Cruz Biotechnology) and HSP90 (1:1000, 610,418, BD Biosciences). Total protein levels were determined with bicinchoninic acid (BioRad, Courtaboeuf, Les Ulis, France) against known concentrations of bovine serum albumin (BSA). Fifty μg of proteins were loaded in each well of 4–12% acrylamide/bis-acrylamide gels ran in denaturing conditions then transferred onto PVDF membranes for immunodetection. HRP activity conjugated to secondary antibodies was revealed with ECL substrate (Amersham, Bethesda, MD, USA). Images were subtracted of background with ImageJ and optical densities of bands normalized to GAPDH and GR.
Animals
Transgenic lines Thy1-YFP (B6.Cg-Tg(Thy1-YFP)HJrs/J, APP/PS1 (B6C3-Tg(APPswe,PSEN1dE9)85Dbo/Mmjax) are from Jackson labs (Bar Harbor, ME, USA), and NR3C1 knockin mutant Ser134Ala/Ser267Ala (B6.Tg(Nr3c1tm2/Jean)/J) was previously described [2]). Mice were housed in groups (2–4/cage) with cotton swabs and igloos for nesting, under a 12 h light/dark cycle (on 7 AM, off 7 PM), at 22–24 °C, 50 ± 5% humidity, and ad-libitum food and water. All efforts were made to minimize animal suffering and reduce their number in each experiment. The starting number of mice used included (n = 19 Nr3c1+/+-thy1-YFP, n = 16 Nr3c1ki/ki-thy1-YFP, n = 26 Nr3c1+/+-thy1-YFP-APP/PS1 and n = 19 Nr3c1ki/ki-thy1-YFP;APP/PS1), which decreased in the later age points due to mortality compounded by genotypes and anesthesia during imaging. Mice were acclimated at least 1 h in testing rooms before behavior. Behavior testing was always done in mornings (8 A.M.–12 P.M.) except for the rotarod training (7 P.M.). Equipment was cleaned thoroughly with 30% ethanol between trials. Longitudinal data were collected at 3, 6 and 9 months in different testing rooms by different experimentalists blinded to the age and genotype of the animals. We used both males and females for behavioral characterization of the Nr3c1ki line. We used males only for all multiparametric experiments in multi-transgenic animals.
Open field
Mice positioned in the center freely explored an arena (50 cm × 50 cm, dim light ~ 50 lux) for 10 min filmed with a webcam. Total distance traveled and time spent in the center (29 cm × 29 cm) were determined with EzTrack (available on Github).
Elevated plus maze
Mice positioned in the center freely explored the arms (50 cm × 20 cm elevated 50 cm above floor, dim light ~ 20 lux) for 5 min filmed with a webcam. The number of entries and time spent in each arm were determined manually.
Rotarod training
Mice were habituated on the non-accelerating rotarod (2 rpm, 1 min followed by 30 s rest, repeated 15 trials) for 2 consecutive days before 2 training sessions each of 15 trials on the accelerating rod (from 2 to 80 rpm reached in 2 min with 1 min rest inter-trial) for 2 consecutive days in the evenings (~ 5 lux). Recall was performed 10 days later for 1 session on the accelerating rod as before.
Novel object recognition
Mice positioned in the center freely explored a L-shaped arena (30 cm × 10 cm, dim light ~ 50 lux) for 10 min filmed with a webcam on day 1 for habituation, with identical objects on each side on day 2, and with one previous (Lego blocks) and one novel object (falcon tube) on each side on day 3. Time spent exploring each side on day 1 and touching the objects on day 2 and 3 were determined manually. Object preference was calculated as ratio of time spent with each object; object memory was calculated as index = (novel − known)/(novel + known).
Three-chamber test
Mice positioned in the center freely explored an arena (60 cm × 41 cm divided in 3 equal chambers with 2 doors in the middle, dim light ~ 50 lux) with empty prisons on each side for 10 min filmed with a webcam on day 1, on day 2 with a same-sex juvenile on one side to determine preference, and on day 3 with the previous juvenile on same side and one novel on the other to determine memory. Time spent touching the empty prisons or the juveniles were determined manually. Social preference was calculated as ratio of time spent with the juvenile over the empty prison; social memory calculated as index = (novel − known)/(novel + known).
Barnes maze
Mice positioned in the center explored an arena (92 cm diameter with 20 holes 5 cm each equally spaced, one of which has the hidden escape box elevated 105 cm above floor, bright light ~ 100 lux) for the time necessary to guide the mouse in the correct hole to spend 2 min in the escape box for habituation on day 1. During acquisition on day 2, mice are given 3 min to find freely the hidden box and reside for 1 min. Mice are place back in homecage for 15 min before next trials (repeat 3 times/day for 5 days). Probe trials were conducted for 90 s on day 6 and 14 in which the hidden box is removed to test for short and long term memory. Number of pokes (errors) in each hole and latency to reach the target hole was measured manually.
Y-maze
Mice positioned in the center freely explored the arms (60 cm × 15 cm) for 5 min filmed with a webcam. We counted manually the number of alternations between consecutive arms.
Thinned skull 2-photon microscopy
Mice were anesthetized with a mix of 0.075 mg/g ketamine and 0.01 mg/g xylazine and lidocaine sprayed atop the skull prior surgery. Skull bone was thinned to transparency using disposable ophthalmic surgical blades (Surgistar, Vista, CA, USA). The scalp is sutured and topped with antibiotic cream to avoid infection between imaging sessions. A detailed map of the pial vasculature and dendritic territories were taken for subsequent relocation as previously described [5]. Only males were used because YFP expression is too bright and diffuse in females interfering with the detection of quantifiable isolated dendrites and spines specifically in aged females.
Open skull 2-photon microscopy
A 3–4 mm craniotomy was prepared over the transcranial imaging zone and the underlying dura was removed and kept in an aqueous environment of HEPES-buffered artificial cerebrospinal fluid (ACSF in mM 120 NaCl, 3.5 KCl, 0.4 KH2PO4, 15 glucose, 1.2 CaCl2, 5 NaHCO3, 1.2 Na2SO4, 20 HEPES, pH = 7.4). The cortex was covered by a thin layer of low-melting agarose 0.8% in ACSF to avoid heartbeat motion artifacts as previously described [2]. Hamilton syringe with a glass pipette were used to deliver 1–2(-nitrophenyl)ethyl(S)AMPA at 5 mM or glutamate at 100 mM (Bio-Techne, France) diluted in ACSF through the agarose bed as previously described [3]. Photolyse parameters were tuned to 720 nm, 0.7 mW for 5 s, and directed in motor cortex specifically at the head of new spines formed after the rotarod training. Images were taken for up to 15 min (n = 19 Nr3c1+/+-thy1-YFP, n = 33 Nr3c1ki/ki-thy1-YFP). Control spines did not receive laser stimulation (n = 13 Nr3c1+/+-thy1-YFP, n = 15 Nr3c1ki/ki-thy1-YFP). Spine enlargement was calculated as the % change of brightness in the head defined as region of interest using ImageJ [2].
Image acquisition
Mice (n = 9 Nr3c1+/+-APP/PS1-thy1-YFP, 8 Nr3c1ki/ki-APP/PS1-thy1-YFP) were injected i.p. with 10 mg/kg Methoxy-Xo4 (Bio-Techne) 48 h prior to imaging at 3, 6 and 9 months of age. Just before acquisition, mice were also injected i.v. with 50 μl dextran 70 Kda (25 mg/ml conjugated with Texas red or FITC, Sigma-Aldrich, St. Louis, MO, USA). Hydrazide-AlexA633 (1 mg/kg, ThermoFisher Scientific) was injected i.v. 24 h prior to imaging to mark arterioles. Images were acquired in the somatosensory cortex of deeply anesthetized mice with a FVMPE RS two-photon microscope (Olympus, Hamburg, Germany) equipped with a 25X, numerical aperture 1.05 water-immersion objective (XLPLN25XWMP2, Olympus) and an InSight X3 femtosecond-pulsed infrared laser (Spectra-Physics, Evry, France) for optimal fluorescence excitation and emission separation. Excitation was at 750 nm for Methoxy-Xo4, 780 nm for FITC, 960 nm for YFP and 1040 nm for Texas Red. Images were taken with a digital zoom of 7.2 at each image session using 0.75 μm step with a scanning dwell time of 2.55 μsec per pixel. Laser power was adjusted with the depth but kept below 30 mW. Each scan stacks consists of images at 512 × 512 pixels resolution. Time-lapse acquisition was done in a smaller field of view with galvanometric scanning mode and conventional raster scanning for blood flow measurements based on the line scanning method [4]. Plasma is fluorescent unlike blood cells that do not uptake dextran dyes permitting identification of their circulation. Fluorescence excitation was delivered by a Lambda LS xenon arc lamp (300 W; Sutter Instruments, Novato, CA, USA) fitted with a fast-rotating filter wheel (27 ms lag) and linked to the stereomicroscope with an optical fiber and 20 × objective. Fluorescence emission was captured with a sCmos camera (C11440 Orca-Flash 4.0, Hamamatsu Photonics, Japan) capturing images at 200 Hz that allowed speed limit up to 9 mm/s which is sufficient for capturing flow in arterioles ~ 10 μm diameter. In a separate experiment, mice (n = 4 Nr3c1+/+-APP/PS1-thy1-YFP, n = 4 Nr3c1ki/ki-APP/PS1-thy1-YFP not trained and n = 6 Nr3c1+/+-APP/PS1-thy1-YFP, n = 6 Nr3c1ki/ki-APP/PS1-thy1-YFP trained) were injected i.p. with corticosterone (15 mg/kg, Sigma-Aldrich) 12 h after the last rotarod training, and immediately after the first imaging of the time-lapse session.
Image analysis
The field of view (200 × 200 × 150 μm) in consecutive images was realigned with RegStack plugin and distances between nearest spines along dendrites and from amyloid plaques measured with ImageJ. Regions of interest were drawn to measure the surface of amyloid deposits. The numbers of blood cells were counted in the amyloid-covered vessels in 0.15 mm3 of field of view and normalized to 1 mm3. The number of axonal dystrophies and dendritic spines were expressed as densities. Dextran-Texas Red filled microcirculation but did not penetrate blood cells permitting identification of their circulation as previously described [4]. The change of flow between imaging sessions was determined only in amyloid-covered vessels. Two or more additions (or eliminations) of spines ≤ 5 μm along a dendrite define a dynamic cluster of formation (or elimination) as previously described [18]. All clear headed-protrusions emanating laterally from the dendritic shaft were counted. Approximately 200 dendritic spines from at least 10–20 dendritic segments were counted per conditions throughout the imaging sessions and averaged per animal. The presence, loss and gain of spines were counted between sessions for each segment and plotted as a function of distance to the nearest spine and amyloid plaque. We verified the distance perimeter from a plaque in 3D. Distance measurement between spines was set at the base of the neck to the base of the next spine following the trace of the dendritic shaft. The proportion of clustered formation (or elimination) equals the number of spines in clusters divided by the total number of new spines added (or eliminated) between imaging sessions.
Simulations of the distance between the nearest spine added (or eliminated) were performed to test if the observed distance is different from chance. For this, one dynamic spine was kept in its fixed position while the other dynamic spines were permutated randomly. This operation was repeated as many times as the number of dynamic spines. For each permutation, one spine of a cluster was randomly re-assigned to all possible spine positions on that dendritic segment keeping the other spine in its fixed observed position. Matlab was used to measure the distance between the clustered spines for each permutation and repeated the process 30,000 times to calculate a 99.9% confidence interval for the probability of clustering as previously described [18]. Averaged values yielded a random distribution of any possible spine clusters in defined dendritic territories from which we calculated the Gaussian best-fit value (Mean ± SEM) to which we compared the observed value. Matlab was also used to simulate the restoration of lost spines at any spine position in the dendrite. We measured the distance between a lost spine and the restored one for each permutation (restoration if < 2 μm, de novo addition if > 2 μm) and repeated the process 10,000 times to calculate a 99.9% confidence interval for the probability of restored lost spines as previously described [40].
Electrophysiology
On the next day of the last training session, coronal slices ≈400 μm of motor cortex were cut with a vibratome and transferred to a temperature controlled (34 ± 1 °C) chamber perfused with oxygenated ACSF (in mM: 127,25 NaCl, 1,75 KCl, 1.25 KH2PO4, 1 MgCl2, 2 CaCl2, 26 NaHCO3, 10 glucose) at a rate of 1 mL/min. Stimulation electrodes were positioned in layer 2/3 ≈500 μm away from recording electrodes. Field potentials (FP) were evoked by stimulation of 0.2 ms at 0.03 Hz. Amplitudes were recorded using single stimuli applied every 30 s for at least 30-min to reach stable baseline. High-frequency stimulation (HFS) consisted in 10 trains of 5 Hz stimuli, each composed of 4 (200 ms) pulses at 100 Hz, repeated 5 times every 10 s. Low-frequency stimulation (LFS) consisted in 2 Hz stimulus for 15 min. Stimulus intensity eliciting 50% of the maximum amplitude was used for all measurements to assess the physiological range of saturation upon repeated protocols as described [46]. The GABAA antagonist Bicuculline methiodide (3.5 mM) was applied at the end of recordings to establish the health of the slices and washout with ASCF.