Glymphatic distribution of CSF-derived apoE into brain is isoform specific and suppressed during sleep deprivation

Male C57BL/6 J and NG2-DsRed (Tg(Cspg4-Ds Red.T1)1Akik/J) mice, which are on C57BL/6 J genetic background, were purchased from Jackson Laboratory (Bar Harbor, ME). In mice older than 6–8 weeks, oligodendrocyte progenitor cells rarely express Ds-Red, which is restricted to peri-vascular smooth muscle cells and pericytes [26, 31]. In these mice, arteries and arterioles are labeled with DsRed, whereas veins are DsRed negative [26]. The AQP4 ko (Aqp4 ^−/− ) mice were obtained from Dr Nagelhus and generated as described [32]. While there was no significant sex differences in glymphatic fluid transport of apoE in young mice, males were used for consistency. The mice were housed in the vivarium facilities at the University of Rochester, School of Medicine and Dentistry. All animal studies were performed according to the NIH guidelines using protocols approved by the University of Rochester Committee on Animal Resources. Mice were housed at a maximum of five per cage, as permitted, depending on the experiment, and maintained on a 12:12 light/dark schedule (6 AM: 6 PM) with food and water ad libitum. Mice were anesthetized with a mixture of ketamine (100 mg/kg) and xylazine (10 mg/kg) by intraperitoneal injection (IP).

The lentivirus (pTANK-CMVie-LckEGFP-WPRE) expressing membrane-bound enhanced green fluorescent protein (EGFP) was designed to carry, in the 5' to 3' direction, a central polypurine tract (cPPT) element, a cytomegalovirus (CMV) immediate early promoter, membrane-bounded EGFP, expressed in tandem with a Woodchuck Hepatitis Virus Posttranscriptional Regulatory Element (WPRE). Virus particles (called lenti-EGFP) pseudo typed with vesicular stomatitis virus G glycoprotein were produced, concentrated by ultracentrifugation, and titrated on 293HEK cells (1.0 x10¹² colony forming unit/ml) [33, 34]. Lentivirus-apoE3 (OHS 5899–202616165; Precision LentiORF.APOE3 with stop codon) was obtained from GE Dharmacon, Inc. (Lafayette, CO, USA). The apoE3 was cloned into LentiORD plasmid under the control of the CMV promoter. The expression cassette contains CMV, apoE3, IRES, EGFP and WPRE. We referred to this apoE3 gene delivery as lenti-apoE3 to compare with the lenti-EGFP particle. The internal ribosome entry site (IRES) provides for bicistronic expression, but the expression of EGFP is low compared to that of apoE3 in this cassette. Thus, to get a good signal of EGFP expression that is similar to that of apoE3 we used a mixture of both lenti-EGFP and lenti-apoE3 to provide a more rigorous test of the transduced cells (EGFP) in the presence of released apoE3 (detected by immunolabeling of the human apoE) in some experiments where stated. Reconstituted human lipidated-apoE2, −apoE3 and -apoE4 were obtained from Dr. David Holtzman. The particle size (10–17 nm) and disc shape were similar to the range of HDL-like particles derived from astrocytes or isolated from human CSF [12, 14, 15, 35, 36], and also similar to the reconstituted apoE [35]. The size and shape of lipoproteins in CSF are heterogeneous and may be species dependent [3, 37]. We used human apoE, the most abundant apolipoprotein in CSF, and our analysis only detects apoE. The apoE particles were purified by gel filtration-FPLC (Fast Protein Liquid Chromatography) to minimize the loss of proteins and antibody affinity column, and the particle size determined by non-denaturing gel electrophoresis followed Western blot analysis [12, 35]. In addition, the particle size of each of the three apoE isoform was similar [35]. ISF clearance of apoE is not affected by its particle size [6]. The particle size is within the size of the gaps (~20 nm) between astrocytic end feet [38, 39]. FITC-tagged apoE was prepared following the manufacturer’s instructions (FITC Protein labeling Kit (F-6434, Life Technologies, Carlsbad, CA, USA). ApoE labeled with Alexa Fluor 647 was prepared following the manufacturer’s instructions (Alexa Fluor 647, Microscale Protein Labeling Kit (A30009), Molecular Probes, Inc., Eugene, OR, USA). Lectin was obtained from Vector Laboratory (Lycopersicon esculentum (tomato); Burlingame CA, USA).

Lentivirus carrying only EGFP or human apoE3 was stereotactically injected into the right lateral ventricle (co-ordinates: anteroposterior, −0.4 mm; mediolateral, +1.0 mm and dorsoventral, −2.3 mm), using a pulled glass pipette connected to a Hamilton syringe that is driven by a microinjector pump controller (Micro 4, World Precision Instruments, Inc., Sarasota, FL), at 1 μL/min for 3 min. At the end of the injection, the pipette was left at the injection site for 10 min to prevent reflux of the injectate along the injection track. It was removed slowly over 2 min. After 1 to 8 weeks post-transduction mice were re-anesthetized and perfusion fixed with 4% paraformaldehyde (PFA). Brains were sectioned horizontal at 100 μm thickness using a vibratome (Vibratome Series 1000). Sections were observed for EGFP expression, and processed for immunolabeling of various proteins, including human apoE.

Immunohistochemistry was performed as reported [40, 41]. Mice were transcardially perfused with ice-cold phosphate buffered saline (PBS, pH 7.4, Sigma-Aldrich, St. Louis,MO,USA) followed by 4% PFA (Sigma-Aldrich). Free floating brain sections (100 μm; horizontal or coronal) were immunolabeled for various specific markers, such as neurons (NeuN) and astrocytes (GFAP). The tissue was blocked with 7% donkey serum for 1 h and incubated with the primary antibodies overnight. The primary antibodies were rabbit anti-human specific anti-apoE (1:200, Abcam Inc., Cambridge, MA; ab52607), mouse anti-aquaporin 1 (1:200, Abcam. Ab9566), mouse anti-GFAP (1:500, Sigma, G3893), chicken anti-NeuN (1:500, Abcam, ab134014), chicken anti-GFP antibody (Novus Biological,Littleton, CO,USA. NB100-1614). Then, Alexa Fluor-conjugated secondary antibodies were added and incubated for 2 h at room temperature (Life Technologies, 1:500). DAPI (Sigma, 1:2000) was used to identify cell nuclei. Immunofluorescence was visualized using a Bio-Rad MRC500 confocal scanning head attached to an inverted microscope (IX81, Olympus, Tokyo, Japan) controlled by Olympus Fluoview 500 software. For analysis of apoE intensities immunohistochemistry was performed on 100 μm vibratome-cut brain slices. Neurons were identified by NeuN labeling, astrocytes by GFAP expression and blood vessels by lectin staining. Arteries were identified by the red NG2-DsRed expression. NIH ImageJ software (1.47v) was used to measure the intensity of apoE immunoreactivity. Background intensities (5 per image) were measured in the parenchyma of the same field and subtracted from apoE intensities. The intensity of the immunolabeled apoE on the vessel wall and around (a 5 μm band around the vessel) arterial vessels or veins were quantified in the brain regions of three mice. For co-localization of apoE and neurons, a total of 30 neurons were quantified per brain and the average determined. The person analyzing the images and data was blinded to the experimental design.

Mice were anesthetized, fixed to a stereotactic frame and the cisterna magna exposed and cannulated using a 30G needle [26, 42]. Fluorescent tracers (5 μL aCSF containing apoE (0.5%) and cascade blue tagged dextran (10 kDa, 1%) as the reference inert molecule that is not significantly taken up by cells, transported across the BBB or degraded) were injected at 1 μl/min using a Hamilton syringe connected to a Micro Syringe pump controller (Micro 4; World Precision Instruments, Inc.). The same amount of apoE was used for each isoform for comparison. We used a small volume and a slow rate of injection (5 μL at 1 μL/min), that does not significantly increase the intracranial pressure [28], thereby minimizing CSF reflux into the cerebral ventricles from the cisterna magna [26]. ApoE4 and dextran (10 kDa) were used to characterize the technique since, in contrast to the dextran, apoE4 is retained within the brain to a greater extent compared to the other isoforms [6]. These tracers should provide a better resolution of CSF flow around arterial vessels. To quantify the glymphatic fluid inflow into brain, ¹²⁵I-apoE, at a low concentration (10 nM, due to the greater resolution of radioactivity analysis), and ¹⁴C-inulin (6 kDa, 1.0 μCi, an inert molecule) were intracisternally injected at 1 μL/min for 5 min and after 30 min the brain removed and prepared for radioactivity analysis (see below).

Lectin was administered during the cardio-perfusion stage of the experiment, i.e., intravascularly. The mice were cardio-perfused with cold PBS, containing lectin (0.02 mg/ml), at 2 ml/min for 10 min. This was then followed by perfusion with PFA (4% in PBS), as described above.

In anesthetized mice, a craniotomy (3 mm in diameter) was made over the sensory motor cortex. The dura was left intact and the craniotomy was covered with aCSF and sealed with a glass coverslip. To visualize the vasculature, 0.1 ml BBB impermeable Texas Red-dextran 70 (MW 70kD; 1% in saline, Invitrogen) was injected intravenously immediately before imaging. A Mai Tai laser (SpectraPhysics) attached to a confocal scanning system (Fluoview 300, Olympus) and an upright microscope (IX51W, Olympus) was used for in vivo imaging as described [26, 27]. A 20X (0.9NA) water immersion lens was used to image the cortex, from the surface to a depth of ~150 μm at every 5 μm z-steps. We used 870 nm as excitation wavelength and emission was collected at 575-645 nm. The cerebral vasculature was imaged at a resolution of 512 × 512px. FITC-apoE3 was administered intracisternally at 1 μL/min for 5 min and images acquired at 15 min after the injection.

Mice were sleep deprived from 6 to 9 AM and used immediately for the experiment at zeitgeber time 3 (ZT3). Sleep deprivation was maintained by using an enriched environment, containing novel objects and nesting material, to facilitate spontaneous exploration by the mice, as reported [43]. Mice were continuously observed to ensure that they were indeed awake and exploring the objects. Mice that were not exploring the objects were lightly prodded or gently handled to encourage their continuous activity. This method was chosen to minimize stress to the mice [43]. Food and water were freely available.

Earlier, it was shown that iodination (¹²⁵I-) of lipidated apoE did not affect its clearance from brain ISF when compared to untagged lipidated apoE [44]. There are differenced in clearance rate between lipid-poor and lipidated apoE [6, 44]. Thus, if there were changes in the apoE lipidated state with iodination then it should have altered its clearance rate. To evaluate solute clearance from the brain, radio-labeled tracers (¹²⁵I-apoE3 (prepared as we reported [6]) and ¹⁴C-inulin (6 kDa, PerkinElmer) were injected stereotaxtically into the left frontal cortex, as we recently reported [27]. Briefly, a stainless steel guide cannula (Plastic One) was implanted stereotaxically into the left frontal cortex of anesthetized mice (2% isoflurane) with the coordinates of the cannula tip at 0.7 mm anterior and 3.0 mm lateral to the bregma, and 1.0 mm below the surface of the brain. Animals were allowed to recover after surgery and the experiments performed 18–24 h after the guide tube cannulation, as reported [45]. Experiments during sleep state and sleep deprivation were performed between ZT3 and ZT4. In each mouse, a small volume of mock CSF (0.5 μL), containing ¹²⁵I-apoE (10 nM) and ¹⁴C-inulin (1.0 μCi), was simultaneously injected (33 GA cannula, Plastic One) into the brain ISF over 5 min. At the end of the experiments (90 min) the brain was removed and prepared for radioactivity analysis. Samples were first counted for gamma radioactivity (¹²⁵I-apoE) using Wallac 1471 Wizard Automatic Gamma Counter) and counts corrected for TCA-precipitability, as reported [45]. Then all samples were solubilized in 0.5 ml tissue solvable (Perkin Elmer) overnight followed by the addition of 5 ml of scintillation cocktail (Packard Ultima Gold). Samples were then analyzed in a liquid scintillation counter for ¹⁴C- radioactivity (LS6500 Multi-purpose Scintillation Counter (Beckman Coulter, GA, USA). Calculations: The percentage of radioactivity remaining in the brain after microinjection was determined as % recovery in brain = 100 x (N_b/N_i) (eq. 1), where, N_b is the radioactivity remaining in the brain at the end of the experiment and N_i is the radioactivity injected into the brain ISF. Total clearance was determined as 100-%recovery. ¹⁴C-inulin was used as an inert polar molecule which is neither transported across the BBB nor significantly retained by brain cells; its clearance provides a measure of the ISF bulk flow/convective flow. To confirm that FITC labeling apoE do not affect apoE clearance, we microinjected FITC-apoE3 or ¹²⁵I-apoE3 intracortically, as above, and after 90 min the apoE3 levels remaining in brain were determined and expressed as a percentage of the injected dose that was cleared from brain, as reported [26, 46]. Levels of FITC-apoE3 were determined fluorometrically, as reported [47] and ¹²⁵I-apoE3 analyzed as we reported [6].

A custom ImageJ plugin program was written to analyze apoE distribution radius as a function of distance from the vessel wall. In these studies, horizontal brain sections (100 μm each) from the cortex were used as there were no regional differences in apoE uptake from CSF. Arteries were identified by red NG2-DsRed (smooth muscles) expression in the NG2 DsRed reporter mice, veins were NG2 DsRed negative and the vasculature was labeled with lectin. The vessel diameter was calculated from the brain photomicrographs containing lectin immunofluorescence. The digital images of photomicrographs were used for the image analysis using imageJ software. The plugin plotted the intensity value of the immunofluorescence with respect to distance when a line was drawn across the blood vessel identified by the lectin immunofluorescence. Based on the change in the intensity value the blood vessel diameter was calculated. The cortical brain region was used since penetrating arteries were analyzed and there were no regional differences. All arterial vessels were analyzed in each brain section by a person blinded to the experimental design. Three horizontal brain sections (sensory-motor cortex) were used per mouse brain and there were 5 to 9 mice per group.

Primary cultures of mouse astrocytes and choroid plexus epithelial cells were performed, as reported [40, 48]. In brief, the choroid plexus from the lateral ventricles and 4^th ventricle were removed from 15 C57BL6 male mice (4–8 weeks old), digested in 0.4% pronase in Hank’s balanced salt solution (HBSS) and cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing fetal bovine serum (FBS), epidermal growth factor (EGF) and antibiotics (penicillin/streptomycin) on collagen coated wells. The purity of the choroid plexus epithelial cells was typically >99%, as determined from AQP1 immunolabeling and insignificant levels of astrocytes (GFAP-positive cells), and neurons (NeuN-positive cells). For cultured astrocytes, neocortical astrocytes were prepared from P1 to P4 C57BL/6 mouse pups, as previously described [40]. Briefly, cerebral cortices were separated and meninges were removed. The cortical tissue was dissected and washed three times in HBSS without Ca²⁺. After trituration and filtering through the 70-μm nylon mesh, the cells were centrifuged at 200 g for 10 min. The pellet was dissociated to single cell suspension in DMEM/F12, containing 1% Penicillin/Streptomycin and 10% FBS, and plated in T25 culture flasks. Typically, cultures with more than 99% of GFAP-positive cells were used in this study. All cells were grown at 37 °C in a humidified incubator (5% CO₂/98% humidity). Medium was changed after 24 h and every 3 days thereafter. Conditioned media from the near confluent cells were collected, complete protease inhibitors (Roche Applied Sciences) added, centrifuged at 2000 g for 10 min to remove cellular debris and stored until required for analysis. Cells were then washed with PBS, harvested, pelleted and sonicated in lysis buffer.

Frozen cerebral cortex was sonicated in eight times volume of ice-cold RIPA lysis buffer, containing complete protease inhibitor (Roche Applied Sciences) and 30 mM β-mercaptoethanol (BME), and centrifuged at 10,000 g for 15 min in a microcentrifuge at 4 °C (Eppendorf). The supernatant was used to determine apoE levels using Western blot analysis. The same amount of total protein was loaded into each well. Protein levels were determined by using a BCA protein assay kit (ThermoFisher, Scientific).

ApoE3 levels were determined by western blot analysis. Ice-cold CSF (10 μL) and standards (apoE3, 0.1 mg/ml) were diluted in an equal volume of Laemmli sample buffer (BioRad Laboratories, Hercules, CA; 161-0737) supplemented with 200 mM dithiothreitol (DTT; Sigma). Equal volumes (20 μL) of CSF and apoE3 standards were added to the sample wells of 4–15% Precast Gels (Mini-Protein TGX, BioRad, 456-0783) for electrophoresis under reducing conditions followed by transfer to a nitrocellulose membrane (BioRad, 162-0115) and detection using human specific anti-apoE antibody (Abcam, Ab52607; 1;1000) and a secondary antibody (goat anti-rabbit IgG-HRP, Santa Cruz sc2030; 1:10,000). For mouse apoE levels in brain a mouse specific anti-apoE with extremely low recognition of human apoE [49] was used (1:500; Ab20874, Abcam). Mouse apoE in the conditioned media was detected with an antibody (Calbiochem; Cat # 178479). Blots were developed using enhanced chemiluminescence (SuperSignal, 34077, Pierce, Rockford, IL). Levels of apoE3 in CSF were determined from the linear part of the standard curve that was generated from the intensities of Western blots of known amounts of apoE3. The relative abundance of human and mouse apoE in brain were determined with reference to β-actin. Conditioned media from the primary cultures of choroid epithelial cells and astrocytes were concentrated by ultrafiltration using a 10 kDa cut-off filter (Amicon Ultra Centrifuge Filters, 10 k NMWL, Merck Millipore, Cork, Ireland). The wells were loaded with equal volume of the concentrated conditioned media, and un-saturated intensities standardized to the protein concentration of the cultured cells.

Human apoE3 levels were determined using a human apoE specific ELISA kit (Abcam, Ab 108813) by following the manufacturer’s instructions. Samples of CSF from controls (Lenti-EGFP) were used to confirm human specificity. This signal was subtracted from the signal obtained in the CSF collected from the lenti-apoE3 transduced mice. Levels of apoE3 were determined from a standard curve.

Data were analyzed using either Student’s t test when comparing two groups or analysis of variance (ANOVA) followed by post hoc Tukey test when comparing more than two groups. The differences were considered to be significant at p < 0.05. All values were expressed as mean ± SEM.