The relatively large hydrated diameter of lipidated apoE particles prohibits collection by commonly used lower MWCO microdialysis probes (typically 6-40 kDa). To circumvent this issue, we optimized the use of a 1,000 kDa MWCO microdialysis probe for collection of apoE. We first characterized the efficiency of apoE collection from human CSF using microdialysis in vitro
. The recovery of analyte by microdialysis is inversely related to the flow rate through the microdialysis probe[10
]. By extrapolating along the recovery curve to a zero flow rate, it is possible to estimate the maximum amount of exchangeable analyte [10
]. Microdialysis samples were collected hourly at flow rates ranging from 0.4 μL/min to 1.6 μL/min. As expected, decreasing the flow rate increased the amount of apoE detected in microdialysis samples (Figure A). We estimated the total amount of exchangeable apoE to be 111.2 ± 14.4 ng/mL, which represented ~2% of the 6024 ± 152.3 ng/mL total apoE detected within the CSF samples.
Figure 1 Analysis of apoE levels by microdialysis in vitro and in vivo. A. Microdialysis samples were collected hourly from human CSF in vitro at flow rates ranging from 0.4 μL/min to 1.6 μL/min. The concentration of apoE within microdialysis samples (more ...)
Having determined it was possible to collect apoE by microdialysis using a 1,000 kDa MWCO probe, we next tested whether we could collect apoE in vivo
. We implanted microdialysis probes into the hippocampus of 3-4 month old wild-type mice and collected microdialysis samples every two hours at a constant flow rate of 1.0 μL/min. Under these conditions, levels of apoE within the hippocampal ISF were stable throughout the 36 hour collection duration (Figure B). Similar to previous studies, we observed only minimal astrogliosis in the cortex proximal to the site of cannula insertion and no evidence of substantial inflammation following microdialysis (data not shown)[14
]. We found the mean concentration of recoverable apoE in the hippocampal ISF to be 10.1 ± 1.8 ng/mL at a constant flow rate of 1.0 μL/min (Figure C). To verify the specificity of our assay we also performed microdialysis in apoE KO mice. As expected, we detected no apoE in the apoE KO mouse ISF dialysate (Figure C). These data confirm that we are able to collect apoE from the brain ISF and specifically measure how it changes over the course of several hours in a single mouse.
Due to the relatively high variance we observed in apoE levels, we performed a power analysis to determine if we could feasibly detect statistically significant differences in ISF apoE levels. Given the observed mean concentration and average standard deviation of murine ISF apoE at a given time point, we estimate ~5-6 animals per group would be required to observe a statistically significant 70% change in apoE levels at an α-value of 0.80 and p-value of 0.05. Furthermore, because in vivo microdialysis enables assessment of changes in apoE levels in a single mouse, normalizing the level of apoE detected to a baseline value (i.e. the mean level from first 9 hours of collection) for each animal eliminates inter-animal variability in apoE levels and allows for detection of a statistically significant 50% change in relative apoE levels with ~5-6 animals.
Recent studies propose that modulating apoE levels and lipidation in the brain using nuclear hormone receptor agonists may be an effective therapeutic target for AD [8
]. We tested whether we could detect changes in ISF apoE and Aβ levels following administration of the retinoid-X-receptor (RXR) agonist bexarotene using microdialysis. We monitored hippocampal ISF apoE and Aβ levels in 2-month old APP/PS1 mice. Following establishment of a 6-hour baseline apoE and Aβ level, bexarotene (100 mg/kg) or vehicle (water) were administered to the mice via oral gavage. Bexarotene treatment led to a steady increase in ISF apoE levels beginning ~12 hours post-administration (Figure A). ISF apoE levels were increased 2.5-fold 30-36 hours post-treatment (Figure A and B). ISF Aβ levels were decreased by ~35%, similar to previous observations (Figure A and B) [8
]. These data demonstrate the utility of microdialysis to detect biologically relevant, pharmacologically-induced changes in ISF apoE levels.
Figure 2 Bexarotene increases ISF apoE levels and decreases ISF Aβ levels. A. ISF Aβx-40 and apoE levels in the hippocampus of 2-month old APP/PS1 mice were monitored using in vivo microdialysis. Following establishment of a 6 hour baseline ISF (more ...)
To further validate our technique we tested whether we could determine an absolute concentration of recoverable apoE within the hippocampal ISF by the extrapolated zero-flow method. We implanted microdialysis probes into the hippocampus of 3-4 month old apoE3 KI mice and collected ISF dialysate at flow rates ranging from 0.4 μL/min to 1.6 μL/min. We then fit a single-exponential decay curve to the concentration of apoE as a function of flow rate (Figure A). Our analysis determined the concentration of recoverable apoE3 in the ISF under steady-state conditions was 38.1 ± 4.0 ng/mL (n=4).
Figure 3 Analysis of hippocampal ISF apoE3 levels and lipidation. A. The apoE concentration in microdialysis samples collected at flow rates ranging from 0.4 μL/min to 1.6 μL/min was determined by ELISA. Data are presented as mean ± SEM (more ...)
ApoE particles in the CSF or secreted by cultured astrocytes are a heterogeneous mixture that vary in size depending upon the degree of lipidation [5
]. We compared the lipidation of apoE3 particles collected from brain parenchymal ISF to apoE3 particles in CSF by non-denaturing gel electrophoresis, which separates proteins based upon their hydrated diameter. As previously reported, apoE3 particles in CSF were heterogeneous in size, ranging from 8.1 nm to 17.0 nm (Figure B) [21
]. ApoE3 particles collected from ISF were lipidated and were similar in size to particles found in CSF (Figure B). We further verified that the lipidated apoE3 particles in ISF samples were not due to contamination from the artificial cerebral spinal fluid (aCSF) used for microdialysis (Figure B).
Previous studies found an isoform-dependent effect on apoE levels in the hippocampus, cortex, and CSF in apoE KI mice with apoE2-expressing mice having the highest levels of apoE and apoE4 expressing mice the lowest [22
]. Therefore, we tested whether we could detect isoform-dependent differences in apoE levels in the hippocampal ISF by microdialysis. We implanted microdialysis probes into the hippocampus of 3-4 month old apoE2, apoE3 or apoE4 KI mice and assessed the absolute concentration of human apoE in the ISF dialysate by ELISA. The concentration of apoE2 within the ISF (46.7 ± 15.6 ng/mL, n=3) was significantly greater than that of apoE4 (12.6 ± 1.8 ng/mL, n=4) and the concentration of apoE3 (18.7 ± 5.3, n=6) was at a level between that of apoE2 and apoE4 (Figure A). To confirm that the observed differences were not due to differential detection of apoE isoforms by the antibodies used in the ELISA, we measured equal amounts of recombinant apoE2, apoE3, and apoE4 by ELISA and found no significant difference in detection among the apoE isoforms (Figure B). Therefore, these results support the utility of our assay for observing differences in apoE levels in the ISF that are similar to those seen in CSF and in the brain.
Figure 4 Isoform-dependent differences in apoE levels in the hippocampal ISF. A. ISF samples from the hippocampus of apoE2 KI, apoE3 KI, and apoE4 KI mice were obtained by microdialysis using a constant flow-rate of 1.0 μL/min. ApoE levels were assessed (more ...)