Neuronal Activity Regulates Blood-Brain Barrier Efflux Transport through Endothelial Circadian Genes

Abstract

The blood vessels in the central nervous system (CNS) have a series of unique properties, termed the blood-brain barrier (BBB), which stringently regulate the entry of molecules into the brain, thus maintaining proper brain homeostasis. We sought to understand whether neuronal activity could regulate BBB properties. Using both chemogenetics and a volitional behavior paradigm, we identified a core set of brain endothelial genes whose expression is regulated by neuronal activity. In particular, neuronal activity regulates BBB efflux transporter expression and function, which is critical for excluding many small lipophilic molecules from the brain parenchyma. Furthermore, we found that neuronal activity regulates the expression of circadian clock genes within brain endothelial cells, which in turn mediate the activity-dependent control of BBB efflux transport. These results have important clinical implications for CNS drug delivery and clearance of CNS waste products, including Aβ, and for understanding how neuronal activity can modulate diurnal processes.

Keywords: Blood-Brain Barrier; Cerebrovascular Function; Circadian Rhythms; Efflux transport; Neuronal Activity; Neurovascular Signaling; P-glycoprotein; Transcriptomics.

Figure 1.. DREADDs as a Tool to Manipulate Glutamatergic Activity in vivo

(A) Schematic representation of the genetic mouse models utilizing DREADDS to activate or silence glutamatergic activity in vivo. (B) Photomerge of a representative sagittal section of a brain derived from an Activating DREADDs mouse stained for rabbit anti-HA (red) and cell nuclei with DAPI (blue). Color-matched boxes were added to the edges to make a rectangular image. Scale bar is 2 mm. (C-D) Power density spectrograms depicting gamma local field potential (LFP) power from representative multielectrode array electrophysiological recording sessions from Activating DREADDs mice and controls (C) and Silencing DREADDs mice and control (D) CNO dosages are noted, and delivered when indicated by the green (C) or red (D) arrow on the x-axis. (E-F) Average gamma local field potential (LFP) power as a percentage of the pre-CNO injection baseline for Activating DREADDs mice and controls (E) and Silencing DREADDs mice and control (F). Data represent mean ± SEM (error bars). n=3 per group, *p<0.05 by unpaired Student’s t-test.

Figure 2.. DREADDs-Mediated, Neuronal Activity-Regulated Brain EC Transcriptome

(A) Representative FACS plot of the gating strategy used to sort brain ECs. First, intact cells were gated using forward and side scatter (top 3 panels). Next, cells were gated against dead cells and (DAPI-positive) and pericytes and immune cells (FITC-positive). Finally, Alexa 647-positive ECs were positively selected. (B) MA plot representing global gene expression changes in brain ECs after glutamatergic activation vs. control. Red dots signify statistically significant changes by Wald Test. n=4 mice per group. (C) MA plot for glutamatergic silencing vs. control. n=4 mice per group. (D) Venn Diagram for statistically significant (p<0.05 by Wald Test) gene expression changes after glutamatergic activation and silencing. “Neuronal activity-dependent genes” are the 243 (105 and 138) that were regulated in opposite directions after glutamatergic activation and silencing. (E) Clustering heat map of a refined list (at least 0.25 log2 fold change in both directions with Abcd4 added for comparison) of statistically significant (by Wald Test) oppositely regulated neuronal activity-dependent genes. Color scale represents arbitrary units of expression. Red represents lower expression and green represents higher expression.

Figure 3.. Neuronal Activity-Regulated BBB Transcriptome

Heat map for binned p-values and activity-regulated directionality of common BBB genes in Activating vs. Control (left) and Silencing vs. Control (right). Genes were divided into groups for different BBB properties: Tight junctions, Slc transporters, Abc transporters, other transporters, transcytosis, Leukocyte Adhesion Molecules (LAMs), or Other BBB-enriched. Color scale denotes if a given gene was upregulated (↑) (red) or downregulated (↓) (blue) and whether the change was statistically significant by Wald Test (intensity of color). Individual p-values are shown for significantly regulated genes.

Figure 4.. Neuronal Activity Regulates ABC transporter expression and function

(A) Log2 fold change of mRNA expression in Activating or Silencing groups relative to respective littermate controls of 5 major BBB-specific ABC transporters after DREADDs-mediated manipulation of glutamatergic activity. Data represent mean ± SEM (error bars). n=4 per group. *p<0.05, **p<0.005, ***p<0.001, n.s. (not significant) by Wald Test. (B) Normalized Rhodamine123 (Rh123) fluorescence of Activating vs. Silencing cortices/hippocampi following CNO administration. Mutants were paired with littermate controls. The rhodamine fluorescence (brain:blood) of each mutant was normalized to the fluorescence of its littermate control (Excitation=505nm, Emission=560nm). Data represent mean ± SEM (error bars). Individual data points are shown. *p=0.0291 by unpaired Student’s t-test. (C) Normalized Daunorubicin (DNR) fluorescence of Activating vs. Silencing cortices/hippocampi following CNO administration. Mutants were paired with littermate controls The DNR fluorescence (brain:blood) of each mutant was normalized to the fluorescence of its littermate control (Excitation=470, Emission=585). Data represent mean ± SEM (error bars). Individual data points are shown. *p=0.0195 by unpaired Student’s t-test. (D) Normalized Sodium Fluorescein (NaFl) fluorescence of Activating vs. Silencing cortices/hippocampi following CNO administration. Mutants were paired with littermate controls. The NaFl fluorescence (brain:blood) of each mutant was normalized to fluorescence of its littermate control (Excitation=480nm, Emission=538nm). Data represent mean ± SEM (error bars). Individual data points are shown. n.s. (not significant) by unpaired Student’s t-test.

Figure 5.. Neuronal activity-regulated PAR bZip transcription factors modulate Pgp expression and function

(A) Log2 fold change of mRNA expression in Activating or Silencing groups relative to respective littermate controls of 3 PAR bZip transcription factors after DREADDs-mediated manipulation of glutamatergic activity. Data represent mean ± SEM (error bars). n=4 mice per group. **p<0.005, ***p<0.001 by Wald Test. (B) Relative mRNA expression of Abcb1a normalized to GAPDH across a 24 hour day (12:12 hour dark:light) in Wildtype (left) and PAR bZip triple knockout mice (right). Expression levels represent 2^−ΔΔct. Data represent mean ± SEM (error bars). n=4 mice per group. (C) Schematic of genetic strategy to modulate circadian gene oscillation exclusively in ECs in response to tamoxifen. Bmal1 floxed mice were mated to VECadherin-Cre^ERT2 mice and injected with tamoxifen at 5 week of age. Ablation of Bmal1, the master regulator of the positive loop will ablate typical expression of PAR bZip transcription factors. (D) Rhodamine123 (Rh123) fluorescence (brain:blood) in littermate controls (left) and EC-Bmal1 knockout mice (right) across a day (Excitation=505nm, Emission=560nm). Data represent mean ± SEM (error bars). n=3–7 mice per group. (E) Sodium Fluorescein (NaFl) fluorescence (brain:blood) in littermate controls (left) and EC-Bmal1 knockout mice (right) across a day (Excitation=480nm, Emission=538nm). Data represent mean ± SEM (error bars). n=3–6 mice per group. Rhythmicity was assessed by linear regression. Results are represented as solid lines if the statistical model indicates rhythmicity. Non-rhythmic fits are represented by dashed lines. The statistical model was only considered if the BICW > 0.4 (B, D, E). (F) Relative mRNA expression of EC Abcb1a normalized to Rps20 in FACS-purified brain ECs from littermate control mice (left) and EC-Bmal1 conditional knockout mice (right). Expression levels represent 2^−ΔΔct. Data represent mean ± SEM (error bars). Individual pairs are also shown (n=6 mice per group). ***p<0.001, n.s. (not significant) by paired Student’s t-test. (G) Box and whisker plot depicting percent change of Abcb1a mRNA expression after kainic acid treatment in littermate controls vs. EC-Bmal1 cKOs. n=6 mice per group. *p<0.05 by unpaired Student’s t-test. Data represent median, upper and lower quartiles, and minimum and maximum.

Figure 6.. Behaviorally motivated, Neuronal Activity-Regulated Brain EC Transcriptome

(A) Representative sections of barrel cortex in whisker-shaven, environmentally-null mice (top) vs. whisker-intact, environmentally-enriched mice (bottom) stained for goat anti-cFos (red). Scale bar is 100 μm. (B) Quantification of cFos+ cells per mm² in barrel cortex of “−Whisker” vs. “+Whisker” mice. Data represent mean ± SEM (error bars). n=3 per group. ***p<0.001 by unpaired Student’s t-test. (C) MA plot for +Whisker vs. −Whisker. Red dots signify statistically significant changes by Wald Test. n=3 mice per group. (D) Heat maps of genes statistically significantly (by Wald Test) regulated by DREADDs-mediated glutamatergic activation viewed in the Activating replicates (Act) and paired littermate control replicates (Con) (top) and the same genes viewed in the +Whisker replicates (+) and −Whisker replicates (−) (bottom). Color scale represents arbitrary units of expression. Blue represents lower expression and red represents higher expression. Pearson Correlation Coefficient shown. (E) X/Y scatter plot of the same genes in (D) in the Activating mice and +Whisker mice based on log2 fold change. Each dot represents an individual gene. The majority of genes cluster in the top right (upregulated in both) or bottom left (downregulated in both) quadrants indicating a close correlation of expression changes of these genes between these data sets. ABC transporter genes (blue) and PAR bZip transcription factor genes (red) are shown as colored dots. (F) Heat maps of genes statistically significantly (by Wald Test) regulated by DREADDs-mediated glutamatergic silencing viewed in the Silencing replicates (Sil) and paired littermate control replicates (Con) (top) and the same genes viewed in the +Whisker replicates (+) and −Whisker replicates (−) (bottom). Color scale represents arbitrary units of expression. Blue represents lower expression and red represents higher expression. Pearson Correlation Coefficient shown. (G) X/Y scatter plot of the same genes in (F) in the Silencing mice and +Whisker mice based on log2 fold change. Each dot represents an individual gene. The genes do not cluster within specific quadrants indicating a lack of correlation of expression changes of these genes between these data sets.

Figure 7.. Model of neuronal activity-dependent expression of EC circadian clock-regulated PAR bZip transcription factors which regulate BBB efflux transport

Mice are more active during the night vs. day and thus exhibit more neuronal activity during the night. The expression of the PAR bZip transcription factors in brain ECs is inversely regulated by the amount of glutamatergic activity in the brain and their expression regulates the expression and function of BBB efflux transporters. Therefore, there is more BBB efflux during the day/rest period vs. the night/active period in mice. This may be important for maintaining neurochemical balance. Disrupting this system via sleep deprivation or dysfunctional EC circadian function could lead to neurochemical imbalance and impaired waste clearance.