Supplemental Information.
Methods
Animals
Studies used young adult male mice that were group housed (≤5 per cage) on a 12 hr on/ 12 hr off light cycle with food and water ad libitum. Experiments were initiated from 8 to 10 AM for electrophysiology and treatments leading to biochemical measures, and from 9AM to 12 PM for behavioral studies.
Hippocampal Slices and Extracellular Field Recordings
Animals used for extracellular field recordings were 5 to 8 weeks old. The preparation of hippocampal slices and field recordings from an ACSF perfused interface recording chamber were as described.1,2 Promptly after sectioning, slices (370 µm thick, horizontal) were transferred onto an interface recording chamber continuously perfused with preheated oxygenated artificial cerebral spinal fluid (ACSF; 31 ± 1oC; 95% O2 / 5% CO2) containing (in mM): 124 NaCl, 3 KCl, 1.25 KH2PO4, 1.5 MgSO4, 26 NaHCO3, 2.5 CaCl2, 10 dextrose and 0.01 picrotoxin (PTX) at a rate of 60-70 ml/hr. Experiments were initiated approximately 1.5 hrs after slices were placed on the recording chamber. Field excitatory postsynaptic potentials (fEPSPs) were recorded using a glass recording electrode (2M NaCL fill; 2-3MΩ) and bipolar stimulating electrode (twisted, 65 µm diameter nichrome wire) as described.2 For perforant path components, evoked responses were initially tested with paired-pulse stimuli to confirm specificity of potentials and thus electrode placement: LPP responses express paired-pulse facilitation and MPP responses show paired-pulse depression.3 Single pulse baseline stimulation was applied at 0.05 Hz with intensity was adjusted to achieve 50-60% of the maximum spike-free fEPSP; baseline responses were collected for at least 20 min before any manipulation. For both LPP and MPP, potentiation was induced using one 100 Hz train lasting 1 s with stimulus duration and intensity increased by 100% and 50% of baseline levels, respectively, and with 10 µM PTX included in the tissue bath. Induction of perforant path LTP is typically accomplished with 100 Hz stimulation,4,5 a frequency in the high gamma range characteristic of hippocampal and entorhinal neuronal firing during behavior.6,7In all instances, initial slopes and amplitudes were measured from digitized fEPSPs (NACGather 2.0, Theta Burst Corp., Irvine, CA). Assessments of the level of potentiation were made by comparing fEPSP slope and amplitude during the last 5 minutes of recording withthe mean response in the last 5 minutes of baseline recording.
Whole-Cell Recordings
Whole-cell recordings used 4-7 MΩ pipettes filled with (in mM): 130 CsMeSO4, 10 CsCl, 8 NaCl, 10 HEPES, 0.2 EGTA, 5 QX-314, Mg-ATP, 0.3 Na-GTP 1. All responses were initially tested with paired-pulse stimuli. EPSCs were recorded by clamping the granule cell at -70 mV in the presence of 50 µM PTX. For the LPP and MPP, potentiation was induced using a pairing protocol of 2 Hz stimulation for 75 s at a holding potential of -10 mV, after recording a stable 10 min baseline. For optogenetic stimulation of the dentate C/A system, hippocampal slices were prepared from mice that had received a unilateral injection of AAV-ChR2 (details below) centered on the DG hilus / field CA3c 4 to 5 weeks earlier. Recording electrodes and conditions were as described above. TheChR2-expressing C/A axons (verified by GFP labeling) were stimulated by blue light pulses generated using a Master-8 stimulator (A.M.P.I.) and a blue LED (460 nm, 320 mW/mm2, Prizmatix Ltd) connected to a 100 μm core optic fiber positioned above the target dentate gyrus.8 Responses were initially tested with paired-pulse stimuli. Potentiation was induced using a pairing protocol of 2 Hz stimulation for 15 s at a holding potential of -10 mV, after recording a stable 10 min baseline.
Drug Application
For hippocampal slice experiments, compounds were introduced to the ACSF perfusion line using a syringe pump set at a rate of 6 ml/hr.1 Field recording studies used APV (100 µM), WIN55,121-2 (5 µM), physostigmine (2 or 10 µM) and CNO (10 µM) from Tocris Bioscience; JZL184 (1 µM, RTI International); PTX (10 µM for all field LTP experiments; Sigma-Aldrich). DL-APV sodium salt, physostigmine, and CNO were dissolved in water. Other compounds were dissolved in 100% dimethyl sulfoxide and diluted in ACSF to a final concentration of ≤ 0.1% diluent in the ACSF bath. For behavioral experiments, mice were given IP injection of CNO (1 or 5 mg/kg in saline), JZL184 (8 mg/kg in in 80% polyethylene glycol 400 (PEG400)/20% Tween-80), or AM251 (3 mg/kg, dissolved in 10% PEG400/10% Tween 80/80% saline) at 30 min, 4 h, and 1 h before behavior, respectively.
Fluorescence Deconvolution Tomography (FDT)
Hippocampal slices of WT and Fmr1-KO mice were immersion fixed in 4% paraformaldehyde and sectioned (20 µm) on a freezing microtome, and then the top 8 sections (from the surface that was uppermost within the chamber) were slide mounted and processed for immunofluorescence9 using cocktails of primary antisera including rabbit antisera to GluN1(#114103, Synaptic Systems),10 GluN2A(#AB1555, EMD Millipore)11 or GluN2B(#AB1557P, EMD Millipore),12 in combination with goat anti-PSD-95 (ab90426, Abcam).13 Secondary antisera from ThermoFisher Scientific included Alexa Fluor 594 anti-rabbit IgG (A21209) and Alexa Fluor 488 anti-goat IgG(A11055) both used at 1:1000 dilutions.
An epifluorescence microscope (Leica DM6000) with a 63x PlanApo objective and CCD camera (ORCA-ER, Hammamatsu) was used to capture image z stacks, through a depth of 2 µm in 0.2 µm z-steps, from the dentate gyrus (DG) molecular layer. Immunolabeling for the excitatory synapses postsynaptic scaffold protein PSD-95 served as a marker for the postsynaptic compartment. The incidence and density of immunolabeling for NMDAR subunits co-localized with PSD-95 were then evaluated using wide field epifluorescence and FDT as described elsewhere.9,14 Briefly, images within each z-stack were deconvolved using iterative deconvolution (99% confidence, Volocity 4.0, PerkinElmer) and then used to construct a 3-dimensional (3D) montage of the sample field. Automated systems9,15 were used to normalize the background density, identify immunolabeled elements within the size and eccentricity constraints of synapses, and quantify (count and measure fluorescence intensity) double-labeling. Elements were considered double-labeled if there was any overlap in the fields labeled with the two fluorophores as assessed in 3D.
Two images were required to sample the full depth DG molecular layer, at a mid-portion of the internal leaf. As fields of the outer, middle, and inner molecular layers (OML, MML and IML, respectively) do not align with the size, or necessarily the angle of captured images, it was necessary to adapt the puncta quantification software to accurately localize labeled elements to the correct layer. To accomplish this, two image z-stacks with 10% overlap were captured and reconstructed in 3D (as above) for a given NMDAR subunit and PSD-95, and also captured in 2D for DAPI. These images needed to clearly contain the DG granule cells and hippocampal fissure. They were then stitched in Fiji using pairwise stitching 16 to calculate their exact overlap. The NMDAR (red) and PSD-95 (green) channels were then run through the puncta quantification software as above. A simulated image of the quantified puncta was created and redundant dots in the overlap zone were eliminated along a line halfway between the two images (to not double-sample any immunolabeled puncta). Next, a line was drawn to identify the upper border of the granule cell layer (and hence, the proximal boundary of the IML). A line was generated orthogonal to this boundary marker (aligned with the proximodistal dendritic axis). Secondary lines orthogonal to the proximodistal marker, and thus parallel to the granule cell boundary marker, were generated 25% along the proximodistal axis to mark the distal limit of the IML, and 45% further distal on the axis to identify the distal boundary of the MML; the remaining distal 30% of the field comprised the OML. These regions were then converted to masks and applied to the puncta counts thus segmenting each large image into three simulated smaller images.
Lipid Quantitation
Levels of 2-AG, anandamide, oleoylethanolamide (OEA) and palmitoylethanolamide (PEA) were determined using liquid chromatography-mass spectrometry (LC-MS) methods as detailed below.
Lipid extraction. Hippocampal slices were quickly frozen upon drug treatments, and stored at −80°C until the lipid analyses. Frozen tissue samples were homogenized in cold methanol (1ml) containing the following internal standards: 2H8-2-AG, 2H4-anandamide, 2H4-OEA and 2H4-PEA for the quantitation of 2-AG, anandamide, OEA and PEA, respectively (Cayman Chemicals, MI, USA). Lipids were extracted with chloroform/methanol/water (2:1:1; v/v/v) and fractionated through open-bed silica column (60-Å 230–400 Mesh ASTM; Whatman, Clifton, NJ) by elution with 1ml of chloroform/methanol (9:1), as previously described.17,18 The eluates were dried under N2 and reconstituted in methanol for LC-MS/MS analyses. Protein concentration was determined in the homogenate to normalize samples using the bicinchinonic acid (BCA) protein assay (Pierce, Rockford, IL, USA).
LC-MS/MS analyses. An Agilent 1200 LC system coupled to a 6410 triple quadrupole MS system was used. Lipids were separated using a XDB Eclipse C18 column (50×4.6 mm i.d., 1.8 μm, Zorbax, Agilent), eluted with a gradient ofmethanolin water (from 90% to 100% in 5 min, to 100% in 7 min and to 90% in 8 min) at a flow rate of 1 ml-min−1. Column temperature was held at 40°C. MS detection was in ESI and positive ionization mode, with capillary voltage at 3.5 kV and fragmentor voltage at 135 V. N2was used as drying gas at a flow rate of 12 L-min−1and a temperature of 350°C. Nebulizer pressure was set at 50 PSI. Quantifications were conducted by isotope dilution, monitoring [M+H]+ in the selected ion-monitoring mode. The multiple reaction transitions monitored were as follows: 2-AG,m/z379→287; 2H8-2-AG, m/z387→295; anandamide,m/z348→62;2H4-anandamide,m/z352→66; PEA,m/z300→62; 2H4-PEA,m/z 304→66; OEA,m/z326→62; 2H4-OEA,m/z 330→66.
Stereotaxic Surgery
Mice were anesthetized with ketamine and xylazine (100 mg/kg ketamine, 10 mg/kg xylazine cocktail, i.p.), given Carprofen (10 mg/kg, s.c.),and then received an intracerebral injection of an AAV construct using a 10 µl syringe equipped with a 33GA metal needle (Hamilton, Reno, NV) and stereotaxic guidance. Designer Receptors Exclusively Activated by Designer Drugs (DREADDs) constructs AAV8-CaMKIIa:HA-hM4Di-IRES-mCitrine or AAV-CaMKIIa-HA-hM3D(Gq)-IRES-mCitrine (University of North Carolina Vector Core), were infused unilaterally or bilaterally into LEC and/or DG using the following coordinates (AP, ML and DV planes, respectively): LEC +1.0, +/-4.0, -4.8 mm from lambda (0.8 ul), DG -2.0, +/-1.7, -2.35 mm from bregma (0.4 µl). The ChR2 construct, AAV5-CaMKII-hChR2 (H134R)-eYFP-WPRE, was infused bilaterally into CA3 -1.8, +/-3.0, -2.4 mm from bregma (0.3 µl). Mice were maintained in their home cages for a minimum of 3 weeks before testing to allow for viral expression.
Odor discrimination behavior
Small molecule odorants (see below) were pipetted onto filter paper into a glass cup (5.25 cm diameter x 5 cm height) with a small (~1.5 cm diameter) hole in the lid to allow the animal to smell the odor. For the ‘habituation’ session, the mouse was placed in the test chamber (30 cm x 25 cm x 21.5 cm plexiglas box) containing 2 cups without odors and allowed to explore for 5 min. After 5 min in an identical (empty) holding chamber, the mouse was returned to the test chamber. For the Serial odor discrimination task, mice then explored cups containing odor A (in duplicate) for 3 min. This was repeated for 3 min trials with cups containing odor B and subsequently odor C. In the final test session the mouse was exposed to different odors: familiar odor A and novel odor D.
Two serial odor control tasks were employed.Control task #1 tested for potential preferences among odors. Mice were exposed to two cups without odor for 5 min, then put in the holding chamber for 5 min (as above). Subsequently, mice were exposed to 1 cup with odor (i.e. A or D) and 1 cup without odor for 3 min and, after 5 min in the holding chamber, were exposed to the odor they did not previously experience (i.e. A or D) and an empty cup. Control task #2 (the ‘2-cue’ task) tested if mice could perform a simple odor discrimination. The paradigm was the same as the serial odor discrimination task, except after exposure to odor A, they were placed in the holding chamber for 21 minutes before exposure to a familiar odor (e.g. odor A) versus a novel odor (e.g. odor D).
Mice were scored as exploring an odor when their nose was within 2 cm and directed towards the odor hole. Mice were excluded from analysis if they did not explore the odors for at least 1 s (total) for training sessions (no mice were excluded based on this criteria), and for at least 1 s per cue during testing, to ensure the mice were exposed to both test odors. Behavioral experiments used a total of 164 mice; eleven mice were excluded based on exploration criteria during the test session. There was no evident genotype effect with regard to exclusions (Fisher’s Test, two tailed; p=0.216), nor were there any significant effects of other variables within specific experiments (i.e., behavioral test, treatment) on exclusions.For each test session a discrimination index was calculated: (t-novel odor) – (t-familiar odor) / (t-both odors) x100, with ‘t’ denoting the time spent exploring. Between subjects, objects and apparatus were cleaned with 70% ethanol. All testing was counterbalanced by location of odors during the test session, genotype, and treatment.
Odorants for behavioral tasks
Odorants used were as follows: odor A: (+)-Limonene (≥97% purity, Sigma-Aldrich); odor B: Cyclohexyl Ethyl Acetate (≥97%, International Flavors & Fragances Inc.); odor C: (+)-Citronellal (~96%, Alfa Aesar); odor D: Octyl Aldehyde (~99%, Acros Organics); odor E Anisole (~99%, Acros Organics) odor F 1-Pentanol (~99%, Acros Organics). These odors were diluted in mineral oil (Rite Aid) and 100 µl was pipetted onto filter paper to reach a final concentration of 0.1 Pa, as WT and Fmr1-KO mice have similar odor sensitivity at this concentration19.
Statistics and Blinding
All results are presented as means ± SEM. Statistical significance (p<0.05) was evaluated using student’s t-test (2-tail unless otherwise noted), the non-parametric Mann-Whitney U test or Wilcoxon signed rank test (for paired comparison of two groups with unequal variance between groups) or the two-way analysis of variance (ANOVA) with a Bonferroni post hoc test. In all cases analyses used Prism software (GraphPad; San Diego, CA) which provides evaluation of the variance within a group and suitability of the test (e.g., parametric or non parametric) for each specific data set.
All analysis of behavioral measures was conducted from video recordings by investigators blind to treatment and genotype. Microscopy analysis was conducted by automated systems and thus blind to treatment group.
Supplementary Figures.
Supplemental Figure 1. SynapticGluN1 and GluN2A but not GluN2B levels are decreased in the middle molecular layer (MML) of Fmr1-KO mice. Plots show the immunofluorescence intensity frequency distributions for synapse-sized clusters of glutamate receptor subunits colocalized with PSD-95 in the dentate gyrus MML. (a) Synaptic GluN1 levels were lower (i.e., the curve was left shifted denoting lower fluorescence intensity) in slices from Fmr1-KOs relative to WTs (p=0.002, F(22,308)=1.96; n=7 WT, n=9 KO).(b) GluN2A levels were lower in KOs vs. WTs (p=0.004, F(26,308)=1.95; n=8 WT, n=7 KO). (c) Synaptic GluN2B levels did not differ with genotype (p=1.00, F(27,513)=0.25, n=10 WT, n=11 KO). 2-way ANOVA.
Supplemental Figure 2. Physostigmine does not influence LPP baseline responses or anandamide (AEA), oleoylethanolamine (OEA) or palmitoylethanolamide (PEA)levels. (a) Infusion of2 μM physostigmine (Physo) for 1 h had no effect on LPP fEPSPs (p=0.18, t(10)=1.27, n=6 for ea). (b-d) Acute hippocampal slices were equilibrated in the recording chamber for 1.5 h and then physostigmine (2 μM) or vehicle was introduced to the ACSF bath perfusate for 1 h before slices were collected for lipid analyses. Physostigmine did not alter whole slice levels of AEA (b, p=0.64, F(3,38)=0.57), OEA (c, p=0.18, F(3,38)=1.71) or PEA (d, p=0.19, F(3,38)=1.68) for either genotype (n=10 for WT+veh, WT+Physo and Fmr1-KO+Physo; n=9 for Fmr1-KO+veh; 2-tail t test for a, one-way ANOVA for b, c and d).
Supplementary Figure 3. Distributions of Gi-DREADD expression and evidence that unilateral Gi-DREADD-mediated silencing had no effect on acquisition in the serial odor recognition task. (a,a’)Images show the distribution of m-citrine expression in a representative case with bilateral injection of the AAV-Gi-DREADD construct in the LEC (LEC injection placement is visible in panel a, at arrow; scale bar = 500 µm). In both the left (a) and right (a’) hemispheres expression is evident in the LPP terminal field in the dentate gyrus molecular layer (arrowheads) as well as within CA3 stratum lacunosum moleculare (arrow, a’); both distributions arise from ipsilateral LEC layer 2 neurons.20 Lesser expression is evident in CA1 stratum lacunosum moleculare on the left (asterisk); this projection arises from layer 3 LEC.(b, b’) Images show granule cell labeling with a representative dentate gyrus DREADD injection; the two images of the same field show (b) DAPI labeling of all cell nuclei in the purple channel and (b’) m-citrine expression from the Gi-DREADD AAV construct in the green channel (scale bar = 50 µm for both). Arrow (in b’) indicatesm-citrine fluorescence in the granule cell layer. (c,c’) Representative images show the distribution m-citrine expression in the left (c) and right (c’) hippocampus from a case with successful unilateral Gi-DREADD injection in the left LEC only (injection site ventral to the plane shown here; scale bar = 500 µm): Note the presence of robust m-citrine fluorescence in the LPP on the left (arrowhead) and the absence of m-citrine fluorescence within hippocampus on the right. (d) Bar graph shows that in the serial odor task, CNO injection prior to training did not impair learning in mice with unilateralexpression of the Gi-DREADD in LEC (unilateral LEC) or dentate gyrus (unilateral DG); similarly, CNO had no effect on learning in mice without DREADD expression (cases with bilateral injection misses). In all three groups, the mice explored the novel odor (D, red bars) more than the familiar odor (A, blue bars) denoting learning. Paired t-tests for D vs A: unilateral LEC (p=0.047, t(6)=2.5; n=7); unilateral DG (*p=0.016, t(5)=3.59; n=6); bilateral miss (**p=0.007, t(9)=3.5; n=10).
References for Supplementary Information
1.Wang W, Trieu BH, Palmer LC, Jia Y, Pham DT, Jung KM, et al. A primary cortical input to hippocampus expresses a pathway-specific and endocannabinoid-dependent form of long-term potentiation. eNeuro 2016; 3.
2.Trieu BH, Kramar EA, Cox CD, Jia Y, Wang W, Gall CM, et al. Pronounced differences in signal processing and synaptic plasticity between piriform-hippocampal network stages: a prominent role for adenosine. J Physiol 2015; 593: 2889-2907.
3.Christie BR, Abraham WC. Differential regulation of paired-pulse plasticity following LTP in the dentate gyrus. Neuroreport 1994; 5: 385-388.