REV-ERB-alpha and -beta coordinately regulate astrocyte reactivity and proteostatic function

Significance

Circadian clock proteins have been implicated in regulation of neuroinflammation and neurodegenerative diseases, though mechanisms are still emerging. Our study shows that REV-ERB-α and -β, which are circadian clock components and nuclear receptors, compensate for one another, and deletion of both unmasks robust inflammatory and proteostatic gene expression in the brain. We find that REV-ERBs regulate astrocyte reactivity state and proteostatic gene expression in a cell-autonomous manner and influence alpha-synuclein pathology. Our findings reveal REV-ERBs as regulators of astrocyte function and potential therapeutic targets for alpha-synucleinopathies.

Abstract

The molecular circadian clock is a ubiquitous transcriptional–translational feedback loop that regulates CNS function, glial responses, and neurodegenerative pathology. The nuclear receptors REV-ERB-α (Nr1d1) and REV-ERB-β (Nr1d2) are components of the core circadian clock which regulate metabolism, neuroinflammatory responses, synaptic pruning, and protein aggregation, though the cell type–specific effects and relative compensatory effects of REV-ERB-α AND -β in the brain are unknown. To study the CNS functions of REV-ERBs, we developed mouse lines with global or astrocyte-specific, conditional knockout of both REV-ERB-α and -β. We demonstrate that inducible postnatal global deletion of both REV-ERB-α and -β unmasks extensive transcriptional changes in the brain in disease-relevant pathways such as protein catabolism, complement, and oxidative stress which are not observed with REV-ERB-α deletion alone, and drives spontaneous astrocyte reactivity. Astrocyte-specific deletion of REV-ERB-α/-β recapitulates this spontaneous astrocyte reactivity phenotype, indicating that REV-ERBs regulate astrocyte activation in a cell-autonomous manner downstream of the core circadian clock. Upstream transcription factor analysis revealed that REV-ERB-α/-β repress transcription of Stat3, and astrocytic deletion of REV-ERBs induced astrocytic STAT3 expression and downstream STAT3-mediated gene expression, providing a mechanistic link to the astrocyte reactivity shift. Dual REV-ERB deletion enhanced astrocyte alpha-synuclein uptake and protein degradation in vitro and mitigated alpha-synuclein spreading pathology in an in vivo model of Parkinson’s Disease. This study reveals REV-ERBs as regulators of astrocyte function and implicates astrocyte REV-ERBs as potential therapeutic targets to prevent synucleinopathies and other neurodegenerative pathologies.

REV-ERB-α and REV-ERB-β are related nuclear receptors that are part of the molecular circadian clock (1). They function as transcriptional repressors (2) and have been implicated the regulation of a diverse and critical set of functions in the central nervous system such as neuroinflammation (3), metabolism (4), and amyloid deposition (5). Because pharmacological agents have been developed that modulate the activity of both REV-ERBs (6, 7), it is important to more carefully explore the pathways that REV-ERBs regulate in the CNS and their therapeutic potential. Of key interest is the role that REV-ERBs play in glia, which exhibit heterogeneous functional states that regulate key processes such as neuroinflammation (8), synaptic homeostasis (9), and proteostasis (10). Both the basal expression and circadian oscillation of REV-ERBs is dependent on BMAL1, the critical transcription factor of the core circadian clock (6). REV-ERBs then feedback to inhibit BMAL1 transcription, serving as an accessory loop of the clock. Previous work by our group has demonstrated that BMAL1 gates astrocyte reactivity in a cell-autonomous manner (11), astrocyte-specific Bmal1 deletion not only suppresses REV-ERB expression, but also drives increased astrocytic protein uptake and catabolism which was protective in in vivo models of Alzheimer’s disease (AD)-related tauopathy and Parkinson’s disease (PD)-related synucleinopathy (12). However, the downstream mechanisms linking the molecular circadian clock to this astrocyte activation phenotype are incompletely understood. Because REV-ERBs are regulated downstream of BMAL1 and are highly potent transcriptional regulators, they could serve as druggable molecular links between the core circadian clock and astrocyte function.

Though REV-ERBs have primarily been investigated regarding their role in peripheral organs (13), there is growing emphasis on REV-ERB regulated pathways in neural cells to modulate CNS pathology (5, 14). Because the new REV-ERB targeting agents affect both REV-ERB isoforms, REV-ERBs need to be studied in tandem. However, most studies of REV-ERB function have only addressed REV-ERB-α alone, omitting the compensatory role REV-ERB-β may play. REV-ERB-α and REV-ERB-β share about 96% homology between their DNA binding domains (2), suggesting a high probability that they share similar DNA binding targets. For example, in the liver there is a high degree of cistromic overlap in REV-ERB-α and -β DNA binding, and dual deletion of REV-ERB-α and -β unmasks a more severe metabolic phenotype (1, 13). Therefore, studies that only target REV-ERB-α overlook the possible masking effect of REV-ERB-β. Additionally, many prior studies utilized whole body germline REV-ERB-α knockout models which have off-target effects due to loss of REV-ERB-α during development. For instance, our group has shown that germline REV-ERB-α knockout causes spontaneous microglial activation and inflammation (3) which is not observed with postnatal microglial deletion (14). Thus, study of postnatal and dual REV-ERB-α/-β deletion in the brain is needed.

Astrocyte reactivity is a multifaceted shift in astrocyte gene expression and function which is highly dependent on stimulus. Astrocytes express various activation markers and transcriptional profiles in response to a variety of perturbations, from acute inflammatory insults like LPS (8) to more chronic neurodegenerative diseases like AD (15) and PD (16). Key transcription factors such as NF-κB (17), NRF2 (18), and STAT3 (19) have been identified that regulate a subset of astrocyte reactivity phenotypes in the context of neurological pathologies. However, few factors have been identified that regulate spontaneous astrocyte reactivity in nonpathological conditions. Given that astrocyte-specific Bmal1 deletion induced a spontaneous protective astrocyte activation phenotype, there must be downstream causal factors and pathways that could give further insight into astrocyte function and avenues for therapeutics. We hypothesize that REV-ERB-α/-β coordinately regulate astrocyte reactivity and proteostasis. Here, we employ inducible, tissue specific REV-ERB-α/-β deletion in vivo in mice and in vitro in primary glial cultures, to examine the role of REV-ERBs in brain health and disease. We have observed that dual deletion of REV-ERB-α/-β reveals profound alterations in brain gene expression, unmasking a key role for REV-ERBs in the regulation of astrocyte reactivity and proteostasis. By understanding REV-ERBs’ role in the brain, and in particular astrocytes, we will gain novel insights on how to best target REV-ERBs therapeutically in the setting of neurodegenerative disease.

Results

Dual Deletion of REV-ERB-α/-β Unmasks Profound Transcriptional Changes in the Brain Not Observed with REV-ERB-α Knockout Alone.

Most prior studies of REV-ERB function in the brain have largely focused on REV-ERB-α, utilizing germline whole-body knockouts of REV-ERBα. However, in the liver, REV-ERB-α and REV-ERB-β share genomic binding sites (2). Thus, these studies do not consider the confounding effects of developmental loss of REV-ERB-α or the potential compensatory effects of REV-ERB-β in masking transcriptional changes after REV-ERB-α deletion. To address these issues in the brain, we generated inducible postnatal whole-body knockouts of REV-ERB-α (CAG-Cre^ERT2; Nr1d1^fl/fl, termed “CAG-αKO”) and of REV-ERB-α/-β (CAG-Cre^ERT2; Nr1d1^fl/fl, Nr1d2^fl/fl, termed “CAG-α/βKO”) (Fig. 1A). These models allow temporal control of directed gene knockout, bypassing potential developmental effects of REV-ERB loss during development. We administered tamoxifen to both CAG-αKO and CAG-α/βKO Cre− and Cre+ littermates at 2 mo of age and harvested them at 8 to 9 for analysis. All cohorts were mixed sex, and while no overt sex differences were observed, the groups were not powered to detect sex differences. All mice were kept under 12 h:12 h light:dark conditions and harvested between ZT6-9 (noon to 3 pm), a time chosen due to peak expression of the REV-ERBα target gene Fabp7. While immunohistochemistry for REV-ERBs in the brain was unreliable in our hands with available antibodies, western blot analysis from the cortex of CAG-α/βKO mice showed qualitative reduction in REV-ERB-α and REV-ERB-β protein levels in Cre+ cortex relative to Cre− control (SI Appendix, Fig. S1). We performed bulk RNA sequencing analysis on cortical tissue to capture a comprehensive list of differentially expressed genes (DEGs) to compare between both REV-ERB models (20, 21). The results revealed more profound transcriptomic changes in the Cre+ CAG-α/βKO than in the Cre+ CAG-αKO compared to their respective Cre− controls. In the CAG-αKO mice, there were minimal genes that crossed the significance threshold (q value < 0.05) with log₂ relative fold change (log₂FC) magnitude greater than 0.5 (Fig. 1B). In striking contrast, there were a much larger number of DEGs in the Cre+ CAG-α/βKO mice relative to their own Cre− control mice that satisfied this significance criteria (Fig. 1C). The differences between the CAG-α/βKO and CAG-αKO mice can be explained by two phenomena. First, we observed genes that had modest expression changes in the CAG-αKO mic that displayed much greater fold changes in the CAG-α/βKO (Fig. 1D). For example, targets directly repressed by REV-ERB-α and -β such as Bmal1 and Fabp7 are increased in the Cre+ CAG-αKO mice compared to controls; however, in the Cre+ CAG-α/βKO mice, the fold change relative to control was substantially greater. Among these was Bag3, a chaperone involved in autophagy which we have previously shown to be increased in Bmal1 KO astrocyte and implicated in regulation of tau and alpha-synuclein degradation (12). We also observed this effect with inflammatory genes such as C3, C4b, Cxcl5, and Ifit3. Second, we observed genes that showed no changes in the Cre+ CAG-αKO that were nevertheless differentially expressed in the Cre+ CAG-α/βKO mice (Fig. 1E). For example, genes involved in oxidative stress response (Prdx6, Nfe2l2, Nqo1, Gsta3, Hmox1), proteostasis (Ctsl), and astrocyte reactivity (Serpina3n, Gfap, Vim, Timp1, Clu, Lcn2) exhibited this effect (SI Appendix, Fig. S2). Notably, Lcn2, which is increased by 30-fold in Cre+ CAG-α/βKO, is a pro-inflammatory Alzheimer’s Disease biomarker which is implicated in mediating both amyloid-β and tau pathology (22–24). Specific DEGs involved in circadian clock, oxidative stress, proteolysis, and complement are shown in Fig. 1F. To evaluate the degree of concordance in gene expression between CAG-αKO and CAG-α/βKO mice, we used more permissive P values (uncorrected P < 0.05 by the 2-tailed test) to allow deeper comparison. Of the 292 transcripts that increased by >twofold in CAG-α/βKO, only 94 (32.2%) were increased by at least 20% in the CAG-αKO (SI Appendix, Fig. S3 A and B). Of 176 genes upregulated by at least 50% in CAG-αKO, 68 of them (38.6%) were increased by at least 20% in CAG-α/βKO (SI Appendix, Fig. S3C). However, this concordance increased with transcripts that showed a higher degree of upregulation. For instance, among the 20 genes which are upregulated by ≥twofold in αKO, 19 of them increase by at least 20% in CAG-α/βKO, and 17 are increased by more than twofold. Thus, a profound transcriptomic shift occurs upon knockout of both REV-ERB-α and -β that is not observed after knockout of REV-ERB-α alone. KEGG pathway analysis of DEGs on CAG-α/βKO mice identified disease-relevant pathways such as phagosome, complement cascade, glutathione metabolism (Fig. 1G, full list in Dataset S1 A–E). GO term analysis using DAVID (25, 26) identified enriched terms related to inflammation (“innate immune response,” inflammatory response,” “cellular response to type II interferon”) as well as terms related to cell adhesion, lipid and glutathione metabolism, circadian rhythm, and multiple signaling cascades (Dataset S1C). Altogether, these results suggest that tandem deletion of REV-ERB-α and -β unmasks profound transcriptional changes related to circadian regulation, inflammation, complement, lipid metabolism, proteostasis, and astrocyte reactivity.

Fig. 1.

Global knockout of REV-ERB-α/-β reveals more profound transcriptomic phenotype than deletion of REV-ERB-α alone. (A) Schematic of inducible global REV-ERB-α (CAG-αKO) and REV-ERB-α/-β knockout (CAG-α/βKO) experiment. Created with BioRender.com/hm4sm29. (B and C) Volcano plots of genes from bulk RNA sequencing of (B) CAG-αKO cortex and (C) CAG-α/βKO cortex with cutoff for significant DEGs at adjusted P value (q) < 0.05. Note that Fabp7 is highly upregulated in the CAG-αKO cortex but its P value is off the scale of the volcano plot. (D) Heatmap of select genes from CAG-αKO and CAG-α/βKO cortex that demonstrate relative fold change which is greater in Cre+ CAG-α/βKO vs. Cre− littermate controls than in the Cre+ CAG-αKO vs. Cre− littermate controls. (E) Heatmap of select genes from CAG-αKO and CAG-α/βKO cortex that have significant differential expression only in the Cre+ CAG-α/βKO condition relative to control. (F) Graphs of selected genes from C and D. Group mean shown with SEM error bars. (G) KEGG pathway analysis of significant DEGs (q < 0.005) from CAG-α/βKO cortex. N = 3 to 6 mice per group. *P < 0.05, *P < 0.01, ***P < 0.001, ****P < 0.0001 by 2-way ANOVA with Tukey’s multiple comparison.

REV-ERB-α/-β Regulate Spontaneous Astrocyte.

Although germline whole body knockout of REV-ERB-α robustly increased markers of astrocytic and microglial reactivity (3) the transcriptomic changes from postnatal deletion of REV-ERB-α alone were not as severe as seen in the embryonic knockout (Fig. 1). Moreover, multiple transcripts involved in astrocyte reactivity, including Gfap, Serpina3n, Clu, Vim, and Lcn2, were increased in CAG-α/βKO mice (Fig. 1E and SI Appendix, Fig. S2). Therefore, we sought to determine whether postnatal REV-ERB knockout can induce astrocyte reactivity. To test this hypothesis, we utilized the previously described global postnatal REV-ERB-α/-β (CAG-α/βKO) knockouts since they demonstrated the most profound transcriptomic differences. We administered tamoxifen to Cre+ and Cre− CAG-α/βKO littermate mice at 2 mo old and aged them to 4 mo old prior to being killed. In the hippocampus, there was a 45% increase in the percent area of GFAP, a pan-reactive marker of astrocyte activation, in the Cre+ mice relative to Cre− controls; in the piriform cortex, there was more than a 20-fold increase in percent area of GFAP in the Cre+ CAG-α/βKO mice (Fig. 2A). Disrupted circadian rhythms could potentially contribute to the astrogliosis phenotype, so we next examined rhythms in behavior via actigraphy under 12 h;12 h light:dark (LD) and constant darkness (DD) conditions. We observed that the diurnal activity period of Cre+ CAG-α/βKO mice is unchanged compared to that of Cre− CAG-α/βKO when kept in standard LD conditions, though these mice do have severely disrupted circadian rhythms in DD (weak, fragmented rhythms with a period ~21 h), in keeping with a previous report (13) (SI Appendix, Fig. S4A). Thus, loss of normal circadian rhythmicity could contribute to the phenotype, though the mice in this experiment were kept under L:D conditions and exhibit grossly normal diurnal rhythms due to masking. Overall, our results indicate a widespread shift in astrocyte reactivity state across multiple brain regions. Additionally, our group has shown that embryonic whole-body REV-ERB-α knockout have profound increases in expression of IBA1, a general microglial marker, and CD68, a microglial lysosomal marker that more closely correlates with phagocytic activation (3). We observed that, in the hippocampus of Cre+ CAG-α/βKO mice, there was 48% increase in percent area of IBA1 relative to Cre− control; in the piriform cortex, there was more than a twofold increase in IBA1 (Fig. 2B). These trends are similar to those of the embryonic REV-ERB-α knockout. To probe the microglial phagocytic activation, we quantified the levels of CD68 that colocalized with IBA1+ microglia. Interestingly, we observed no differences in percent volume colocalization of CD68 with IBA1 (Fig. 2C), a notable distinction from our previous embryonic knockout. Furthermore, we examined a panel of transcripts related to microglial reactivity in our bulk RNAseq data from Cre+ CAG-α/βKO mice, and saw no significant changes compared to Cre− controls (Fig. 2D). These findings suggest that REV-ERB-α/-β deletion can induce increases in IBA1 but do not cause the robust microglial reactivity phenotype we previously noted in germline REV-ERB-α knockouts. Because REV-ERB effects on microglia have been addressed elsewhere (5, 14), we chose to focus on astrocytes in subsequent experiments.

Fig. 2.

Global knockout of REV-ERB-α/-β increases markers of astrocyte and microglial reactivity. (A) Percent area of astrocyte activation marker GFAP is increased in the hippocampus and piriform cortex in Cre+ CAG-α/βKO mice relative to control. (B) Percent area of general microglial marker IBA1 is increased in the hippocampus and piriform cortex. Fig. 1 B/C Insets are representative 40× maximum intensity projections. (Wide image Scale bar, 100 µm; Inset Scale bar, 50 µm.) (C) Imaris 3D volume reconstruction showing there is no change in percent volume colocalization of IBA1 and CD68 in the Cre+ CAG-α/βKO piriform cortex, demonstrating there is no change in microglial phagocytic activation. Representative 3D volume reconstruction of IBA1, colocalized CD68, and merged channels show. (Scale bar, 50 µm.) (D) Heatmap from bulk RNAseq dataset showing no change in a panel of microglial reactivity transcripts in Cre+ CAG-α/βKO mice relative to Cre− controls. Data are normalized to Cre− control for each gene and each column indicates one mouse. In panels A–C, graphs show group mean with SEM error bars. N = 6 to 9 mice per group. *P < 0.05, **P < 0.01, ***P < 0.001 by the unpaired 2-tail t test with Welch’s correction.

REV-ERB-α/-β Regulate Astrocyte Activation in a Cell-autonomous Manner.

While GFAP expression increased in germ-line REV-ERB-α KO mice (3), this occurred in the setting of widespread inflammation and microglial activation, so it was unclear if this was a secondary response. Because astrocytic Bmal1 gates astrocyte activation (11), we hypothesized that REV-ERB-α/-β may be the downstream mediators of this cell autonomous astrocyte reactivity phenotype. To investigate REV-ERB-α/-β specific function in astrocytes, we generated astrocyte specific REV-ERB-α/-β inducible double knockout mice (Aldh1l1-Cre^ERT2; Nr1d1^fl/fl, Nr1d2^fl/fl termed “Aldh1l1-α/β KO”). We administered tamoxifen at 2 mo of age to Cre+ and Cre− to induce knockout in Cre+ mice and aged them to 4 mo old prior to being killed for immunohistochemical and transcriptomic analysis (Fig. 3A). We observed an 80% increase in percent area of GFAP positive astrocytes in the hippocampus and a ninefold increase in GFAP in the piriform cortex in the Cre+ Aldh1l1-α/β KO relative to Cre− control, with an anatomical distribution similar to the CAG-α/βKO mice (Fig. 3B). This phenotype was not due to overt loss of circadian behavioral rhythms, as the Cre+ Aldh1l1-α/β KO mice had the same activity period as the Cre− Aldh1l1-α/βKO in both LD and DD conditions (SI Appendix, Fig. S4B), though effects of changes in cell-autonomous rhythms within astrocytes cannot be excluded. To assay the transcriptomic changes in the brain after astrocytic-specific REV-ERB-α/-β deletion, we performed bulk RNA sequencing analysis on cortical tissue from Cre+ and Cre− Aldh1l1-α/β KO mice. We observed gene expression changes in direct REV-ERB targets consistent with loss of REV-ERB-α/-β mediated repression. For example, we saw a 1.5-fold increase in Arntl, a 1.3-fold increase in Dbp, and more than an eightfold increase in Fabp7 expression (Fig. 3C). While Fabp7 is a known REV-ERBα target (27, 28), our data show that REV-ERBs exert cell-autonomous regulation of Fabp7 in astrocytes. We observed no increases in IBA1 immunoreactivity, IBA1-colocalized CD68 expression, or microglial reactivity gene expression in Cre+ Aldh1l1-α/βKO piriform cortex, as compared to Cre+ controls (SI Appendix, Fig. S5), suggesting that astrocyte-specific REV-ERB α/β KO does not induce robust microglial reactivity. However, many genes which were significantly altered in the CAG-α/βKO cortex were also altered in the Aldh1l1-α/βKO cortex. For example, there was increased expression of protein catabolism genes Bag3 (1.5-fold) and Ctsl (1.25-fold) in the Cre+ Aldh1l1-α/βKO mice relative to control. We also again saw increases in oxidative stress response gene Nfe2l2 (1.2-fold) and complement component C4b (1.5-fold) (Fig. 3D), genes known to be highly enriched in astrocytes. To gain further insight about the pathways that astrocyte REV-ERB-α/-β are regulating, we performed ORA using the DEG set from the Aldh1l1-α/βKO cortex. Notable pathways differentially expressed in the Cre+ Aldh1l1-α/βKO mice included cytokine and cell–cell communication pathways, complement cascade, and amino acid metabolism (Fig. 3E). Further analysis using DAVID (25, 26) identified GO terms related to fatty acid synthesis and circadian rhythms, and KEGG pathways related to PPAR and Hippo signaling, unsaturated fatty acid biosynthesis, and glutathione metabolism (Dataset S2 A–D). We compared the DEGs from our CAG-α/βKO RNAseq data (n = 1,211 DEGs, see Fig. 1) and the Aldh1l1-α/βKO mice (n = 159 DEGs), and found that 92 genes overlapped between the two models (57.9% of DEGs from Aldh1l1-α/βKO) (SI Appendix, Fig. S6). KEGG pathway analysis of these 92 genes again showed enrichment for pathways involved in glutathione metabolism, PPAR signaling, circadian rhythms, fatty acid biosynthesis, and metabolism (Dataset S2C). These biological processes are critical in astrocytes and help maintain brain homeostasis and signify how astrocytic REV-ERB-α/-β regulate brain homeostasis as a whole.

Fig. 3.

Astrocyte-specific knockout of REV-ERB-α/-β increases markers of astrocyte activation. (A) Schematic of 4-mo. old inducible astrocyte-specific REV-ERB-α/-β knockout (Aldh1l1-α/βKO) cohort. Created with BioRender.com/hm4sm29. (B) Percent area of astrocyte activation marker GFAP is increased in the hippocampus and piriform cortex in Cre+ Aldh1l1-α/βKO mice relative to Cre− control. Insets are representative 40x maximum intensity projections. (Wide image Scale bar, 100 µm; Inset Scale bar, 50 µm.) (C) Heatmap of the relative fold change of select genes from bulk RNA sequencing of Cre− and Cre+ Aldh1l1-α/βKO cortex. (D) Select astrocyte-related gene expression increased in Cre+ Aldh1l1-α/βKO relative to control. (E) KEGG overrepresentation analysis (ORA) of gene expression in Aldh1l1-α/βKO mice shows altered pathways related to cell-signaling, complement, and metabolism. Graphs show group mean with SEM error bars. N = 4 to 7 mice per group. *P < 0.05, **P < 0.01, ***P < 0.001 by the Mann–Whitney U-test.

REV-ERB-α/-β Regulate STAT3 Expression and Activity in Astrocytes.

REV-ERB-α/-β function as nuclear receptors with the ability to regulate the expression of other transcription factors, allowing REV-ERBs to expand their regulation of various cellular processes. Therefore, we performed in silico transcription factor prediction analysis on the DEG set from the 8-mo. CAG-α/βKO mice (Fig. 1) to identify candidate transcription factors regulated by REV-ERB-α/-β in the brain. We used the EnsembleTFpredictor tool, which ranks multiple transcription factor predictor tools such as MORA, HOMER, LISA, and BART by prediction confidence (29). MORA and HOMER utilize known DNA-binding motif-based enrichment analysis, while LISA and BART use published ChIP-Seq databases from multiple tissue types. Positive prediction with at least two tools, especially if including both types of modalities, indicate a significant likelihood that these predicted transcription factors explain a subset of the DEGs. As validation of the model’s predictive power, Nr1d1 (REV-ERB-α), Nr1d2 (REV-ERB-β), and Arntl (BMAL1) were all predicted by at least two tools to regulate a set of the DEGs, as one would expect with loss of direct REV-ERB transcriptional repression (Fig. 4A). Furthermore, Rela/p65 of the NF–KB complex was another predicted transcription factor, which our group has previously identified as regulated by REV-ERB-α in microglia (3). Interestingly, STAT3 was predicted with high confidence by the model using three out of the four tools (Fig. 4A). STAT3 is activated by phosphorylation by the JAK2 receptor and then translocates to the nucleus where it drives transcription of its target genes. In the brain, STAT3 is a cell-autonomous regulator of astrocyte activation, and increased STAT3 activity in astrocytes was sufficient to induce spontaneous astrocyte activation (30). To investigate whether REV-ERBs directly regulate Stat3 expression in the brain; we utilized a published ChIP-Seq dataset of REV-ERB-α DNA binding in two different brain regions (4). We observed that REV-ERB-α exhibited strong binding peak enrichment in the promoter region of the Stat3 gene, indicating that REV-ERBs can directly bind at the Stat3 locus and presumably suppress transcription (Fig. 4B). Furthermore, there was an increase in Stat3 expression in the cortex of Cre+ CAG-α/βKO mice relative to Cre− controls, demonstrating that deletion of REV-ERBs derepressed Stat3 expression (Fig. 4C).

Fig. 4.

Global and astrocyte-specific REV-ERB-α/-β deletion increases astrocytic STAT3 and functional STAT3 activation. (A) Select transcription factors positively predicted by at least two in silico tools to regulate DEGs in the 8-mo old Cre+ CAG-α/βKO cortex from Fig. 1. (B) Data abstracted from Adlanmerini et al. (4) demonstrating that REV-ERB-α binds to the promoter region of Stat3 in 2 different brain regions, relative to IgG control. (C) Stat3 gene expression is increased in the cortex of 8 mo. Cre+ CAG-α/βKO mice relative to control. (D) Global REV-ERB-α/-β deletion increases GFAP-STAT3 colocalization in the hippocampus of 4-mo. old Cre+ CAG-α/βKO mice. Imaris 3D volume reconstruction of GFAP, colocalized STAT3, and the merged channels. (E) Western blot analysis shows increased phosphorylated STAT3 (pSTAT3) expression relative to total STAT3α and actin in Cre+ gDKO bulk hippocampal tissue. (F) Astrocyte-specific REV-ERB-α/-β deletion increases GFAP-STAT3 colocalization in the hippocampus of 4-mo-old Cre+ Aldh1l1-α/βKO mice. Imaris 3D volume reconstruction of GFAP, colocalized STAT3, and the merged channels. (G) Western blot analysis shows increased phosphorylated STAT3 (pSTAT3) expression relative to total STAT3α and actin in Cre+ Aldh1l1-α/βKO bulk hippocampal tissue. All graphs are group averages with SEM error bars. (Scale bars, 50 μm.) N = 6 to 9 mice per group for C, 5 to 7 mice per group for F, 4 mice per group for E and G. *P < 0.05, **P < 0.01 by the unpaired 2-tail t test with Welch’s correction.

Given that STAT3 is important for regulating astrocyte reactivity, we next measured STAT3 levels in astrocytes in vivo. We observed an increase in the percent colocalization of STAT3 with GFAP positive astrocytes in the hippocampus in Cre+ CAG-α/βKO mice by immunohistochemistry, indicating an upregulation of STAT3 levels in astrocytes (Fig. 4D). Because phosphorylated STAT3 (pSTAT3) is the active form of the transcription factor, we sought to determine if STAT3 was activated in the Cre+ CAG-α/βKO mice relative to control. In bulk hippocampal tissue, we observed no change in total STAT3 protein levels by Western blot; however, there was a significant increase in pSTAT3 levels as well as the ratio of pSTAT3 to STAT3 expression in the Cre+ CAG-α/βKO mice (Fig. 4E). As reported by other groups, phosphorylated STAT3 was undetectable by immunofluorescence, so we interpret the increased pSTAT3 expression by western blot as evidence of activation of the STAT3 pathway (19). These results suggest that there is increased activity of the STAT3 pathway upon REV-ERB-α/-β deletion and provides in vivo confirmation of the in silico prediction. We next hypothesized that there would be a similar increase in STAT3 expression in the astrocytes in the Cre+ Aldh1l1-α/βKO relative to control. We also observed increased colocalization of STAT3 with GFAP in the hippocampus of Cre+ Aldh1l1-α/βKO mice (Fig. 4F). In bulk hippocampal tissue from Aldh1l1-α/βKO mice, we observed an increase in pSTAT3 protein expression as well as pSTAT3 to STAT3 ratio by western blot in the Cre+ Aldh1l1-α/βKO mice relative to control (Fig. 4G). These results together demonstrate that REV-ERB-α/-β regulate STAT3 expression in astrocytes in a cell-autonomous manner, and that increased STAT3 activity could be regulating the increased astrocyte activation phenotype.

Global REV-ERB-α/β Knockout Reduced the Spread of Alpha-Synuclein Pathology in PFF Injection Model.

Our previous work has shown that global Bmal1 deletion can induce astrocyte Bag3 expression, augment astrocyte proteolytic pathways, and reduce alpha-synuclein pathology in a PFF injection model of synucleinopathy (12). We observed increased expression of genes associated with neurodegenerative disease-associated protein catabolism such as Bag3 and Ctsl in the our REV-ERB CAG-α/βKO mice (Fig. 1), suggesting a potential neuroprotective role for REV-ERBs in neurodegenerative models. Therefore, we utilized the alpha-synuclein PFF injection model, in which exogenous alpha synuclein induces seeding and spreading of endogenous synuclein aggregation (31). To test the hypothesis that global REV-ERB-α/-β knockout could be protective, we administered tamoxifen to 2-mo-old Cre+ and Cre− CAG-α/βKO littermate controls and then waited 1 mo prior to unilateral intracerebral injections of synuclein PFFs into the striatum (Fig. 5A). Three months after PFF injection, we harvested mice and quantified phosphorylated alpha-synuclein pathology in brain regions ipsilateral and contralateral to the injection, in Cre+ CAG-α/βKO mice and Cre− littermates. While seeding of pathology was identical on the ipsilateral side (SI Appendix, Fig. S7), we observed a marked decrease in phospho-synuclein pathology in the contralateral piriform cortex, which was a region with a robust increase in GFAP positive activated astrocytes following REV-ERB-α/-β knockout (Fig. 5B). This result demonstrates that REV-ERB CAG-α/βKO mitigates alpha-synuclein spreading in vivo.

Fig. 5.

REV-ERB-α/β regulate protein degradation in in vivo and in vitro models of alpha-synucleinopathy. (A) Schematic of synuclein preformed fibril (PFF) intrastriatal injection in Cre− and Cre+ CAG-α/βKO mice. (B). Global REV-ERB-α/-β deletion increases percent area of astrocyte activation marker GFAP and decreases percent area of phosphorylated alpha-synuclein (pSyn) pathology in the contralateral piriform cortex. (C) Schematic of in vitro experiment in primary wild-type astrocytes for synuclein PFF uptake and degradation with nontargeting control siRNA (siSCR) or siRNA targeting Nr1d1 and Nr1d2. (D) Representative Western blot showing individual protein knockdown of REV-ERB-α and REV-ERB-β by siRNAs. (E) Representative graph showing Nr1d1 and Nr1d2 double-knockdown astrocytes (siNr1d1/2) display increased mean fluorescent intensity (MFI) of synuclein-PFF uptake (PFF-488) and lysosomal internalization (PFF-pHrodo) than siSCR controls as measured by flow cytometry. All graphs are group averages with SEM error bars. N = 9 to 12 mice for B, n = 3 technical well replicates for D. P values shown are from 2-tailed t tests. A and C were created with BioRender.com/hm4sm29.

Given that the reduction in synuclein pathology in the Bmal1 knockout model was largely driven by astrocyte-specific Bmal1 knockout (12), we hypothesized that astrocyte REV-ERBs could mediate a similar effect. To determine if targeting REV-ERBs in astrocytes could increase protein uptake and degradation, we utilized an in vitro primary astrocyte system (Fig. 5C). We successfully knocked down both REV-ERB-α and REV-ERB-β levels in astrocytes with siRNA as measured by Western blotting (Fig. 5D) and by qPCR (SI Appendix, Fig. S8A). We then incubated cells with fluorescently labeled proteins and measured the intracellular signal (Fig. 5E). To assay protein degradation, we utilized alpha-synuclein PFFs conjugated to pHrodo, a pH sensitive dye that fluoresces only when in acidic compartments such as mature lysosomes. We observed an increase in intracellular PFF and pHrodo signal in REV-ERB-α/-β knockdown astrocytes, as compared to those treated with nontargeting siRNA, indicating increased uptake and lysosomal degradation of PFFs (Fig. 5E). As further confirmation of the lysosomal localization, the PFF-pHrodo signal was nearly completely ablated by treatment with chloroquine (SI Appendix, Fig. S8B), which disrupts the acidification of lysosomes and inhibits autophagy. These results together suggest that knocking out REV-ERB-α/-β in the brain increases protein degradation of pathological alpha-synuclein protein. Furthermore, in vitro studies suggest that REV-ERB-α/-β may be mediating this effect through astrocytes by increasing phagocytic and autophagic activity.

Discussion

New pharmacological agents that modulate REV-ERB-α/-β function are being developed to target a wide variety of diseases, including cancer (32), steatohepatitis (33), and multiple sclerosis (34). To investigate these agents for therapy in neurological diseases, a comprehensive understanding of the role of REV-ERB-α/-β in the brain must be elucidated. Our study investigates REV-ERB-α/-β function in the whole brain without the confounds of embryonic REV-ERB-α knockout as well as the compensation of REV-ERB-β. Here, we show that REV-ERB-α and -β coordinately regulate a larger and more varied transcriptome in the brain than REV-ERB-α alone. This result aligns with previous findings showing that knocking down both REV-ERB-α and REV-ERB-β exacerbated hepatic metabolic gene expression and triglyceride accumulation more severely than knocking down either alone (1), indicating that REV-ERB-β compensated for the loss of REV-ERB-α and vice-versa. One possible explanation of our findings is that a similar regulatory mechanism occurs in the brain where REV-ERB-β masks many of the transcriptomic changes induced by single REV-ERB-α deletion, though we did not test this directly. Additionally, because the postnatal REV-ERB-α deletion had a much less pronounced phenotype than the embryonic REV-ERB-α knockout (3), future studies should utilize these conditional knockout models to avoid the confounding developmental effects of the embryonic knockout. Because existing REV-ERB modulating drugs affect both REV-ERB-α and REV-ERB-β², studies that solely focus on REV-ERB-α to identify candidate druggable pathways omit many critical REV-ERB functions. Our study shows that REV-ERB-α/-β regulates key CNS pathways such as protein degradation, oxidative stress, complement activation, and glial reactivity, all hallmarks of neurodegenerative disease (18, 35–37). REV-ERB-α has been linked to inflammatory modulation (3, 38), oxidative stress regulation (39, 40), and protein clearance (5), though many of these studies utilized germline REV-ERB-α models to study the effects on peripheral tissues or the CNS. Notably, global deletion of both REV-ERB-α and -β unleashes a neuroinflammatory response, including 30-fold induction of the pro-inflammatory factor and Alzheimer’s biomarker Lcn2 (22–24), that is not observed with REV-ERB-α knockout alone, or with astrocyte-specific deletion of REV-ERB-α/β. This inflammatory response could be mediated by loss of REV-ERBs in neurons, other brain cells, peripheral cells/organs, or changes in overall circadian function, and suggests that total loss of REV-ERB-α/β function could have deleterious effects which might be avoided by incomplete or cell-targeted REV-ERB inhibition.

Our group has previously shown that Bmal1, the master regulator of the circadian clock, regulates astrocyte reactivity in a cell autonomous manner (11, 41). This phenotype was striking because this single gene knockout model induced widespread spontaneous astrocyte reactivity. Regarding therapeutic relevance, this Bmal1 KO astrocyte reactivity phenotype was protective in models of tauopathy and alpha-synucleinopathy by increasing astrocyte-mediated protein degradation in vivo and in vitro (12). Therefore, understanding the mechanisms of the circadian clock’s regulation of astrocytes is highly important to leveraging astrocytes to treat neurological diseases. However, targeting Bmal1 directly could prove deleterious, as Bmal1 deletion in astrocytes abrogates the cellular molecular clock, rendering the cell arrhythmic (11). Furthermore, broader Bmal1 deletion alters behavioral rhythms (42), impairs metabolic homeostasis (43), and disrupts astrocytic synaptic regulation via diminished neurotransmitter uptake (44). Therefore, identifying the downstream mechanisms by which the molecular clock regulates astrocyte reactivity would be highly beneficial. We hypothesized that REV-ERB-α/-β could regulate the phenotype because REV-ERB-α/-β is downstream target of BMAL1 transcription and whole-body germline REV-ERB-α knockout increased astrocyte activation (3). However, this astrocyte reactivity phenotype could be a response to developmental effects and resulting neuronal abnormalities. Here, utilizing the global postnatal REV-ERB-α/-β knockout model, we have demonstrated that REV-ERB-α/-β does in fact regulate astrocyte activation, potentially downstream of BMAL1, as many of the DEGs in the global postnatal BMAL1 knockout are also DEGs in the global postnatal REV-ERB-α/-β knockout. We have observed that astrocyte reactivity phenotype was a cell-autonomous effect replicated in astrocyte-specific REV-ERB-α/-β knockout mice. These astrocyte-specific REV-ERB-α/-β knockout mice do not have altered activity periods in the normal light–dark conditions or in constant darkness, so it is unlikely that this effect is mediated by circadian arrhythmicity or desynchrony, though an effect of cell-autonomous rhythm changes within astrocytes cannot be excluded. Thus, these findings point to a cellular mechanism occurring within astrocytes. Previously, we observed that astrocyte specific-BMAL1 knockout was protective in the alpha-synuclein PFF model (12) and here we show evidence that globally targeting REV-ERB-α/-β partially recapitulates the phenotype by reducing spreading alpha-synuclein pathology to the contralateral cortex in vivo. Of note, we did not observe reduction in pathology in the ipsilateral cortex in REV-ERB-α/β DKO mice as we did previously in Bmal1 KO (12), suggest that there may be some accessory REV-ERB-independent pathways by which BMAL1 impacts protein catabolism. While REV-ERBα DKO yields a qualitatively less severe astrocyte reactivity phenotype than we have previously shown with similar Bmal1 deletion (12), this could be a technical issue (related to excising two floxed alleles instead of one), or it could suggest that BMAL1 has additional function. However, reactive astrocytes in the REV-ERB-α/-β knockout condition can still reduce trans-synaptic pathologic spread (45) across multiple synapses, suggesting therapeutic potential. Our in vitro results demonstrate increased uptake and lysosomal localization of synuclein PFFs following REV-ERB-α/-β knockdown in primary astrocytes, suggesting that astrocytes may mediate a significant portion of the protective effect in the in vivo PFF injection model. These findings echo our previous work showing that Bmal1 deficient astrocytes increased pathological protein uptake (12) and autophagic flux (46), again suggesting that REV-ERB-α/-β downstream of BMAL1 mediate these effects.

To gain mechanistic insight how REV-ERB-α/-β regulate the CNS transcriptome and explain the glial reactivity phenotype, we utilized in silico transcription factor prediction analysis (29) of the DEG list in the global REV-ERB-α/-β knockouts. One notable hit was STAT3, a potent transcription factor that regulates cell-autonomous astrocyte reactivity (19, 30). Abjean et al. showed that expression of a constitutively active JAK2 in astrocytes that will continuously activates STAT3 increases astrocyte reactivity and reduces huntingtin protein aggregation in a model of Huntington’s disease (47). Thus, the STAT3 pathway could serve as the mechanistic link between the molecular clock, REV-ERB-α/-β, and astrocyte reactivity. However, STAT3 activation can also exacerbate some pathologies, including amyloid plaque deposition (30, 48), so the effects of REV-ERB deletion will need to be tested in various models in addition to alpha-synuclein. We noted that REV-ERB-α binds the Stat3 promoter in brain tissue (4), and that global loss of REV-ERB-α/β leads to increased Stat3 expression in the cortex. We propose that this gene expression increase is due to the loss of REV-ERB-α/β mediated repression and that thereby leads to increased STAT3 translation and subsequent phosphorylation. As further validation of the upregulation of STAT3 activity in astrocytes, we saw increased STAT3 protein expression in astrocytes in both our global and astrocyte-specific REV-ERB-α/-β knockouts, as well as increased activated phosphorylated STAT3 in whole tissue, demonstrating that the transcriptomic changes are converted into functional protein changes. Previous studies have linked REV-ERBs to STAT3; for example, overexpressing Nr1d1 increased SOCS3, an inhibitor of the JAK/STAT3 signaling, in an ovarian cancer cell line and reduced cell proliferation (49). On the other hand, one group observed that knocking down REV-ERB-α in astrocytes in vitro increased select markers of astrocyte reactivity and decreased neurotrophic support for dopaminergic neurons but decreased STAT3 phosphorylation and activation (50). However, these results are from in vitro studies of astrocytes and do not address the effects of superphysiological upregulation of Nr1d2 due to Nr1d1 downregulation on STAT3 activation. Nr1d2 upregulation could drive upregulation of SOCS3 and thus inhibit STAT3 activity. Alternatively, the STAT3 activity in astrocytes could be driven not by transcriptional regulation of Stat3 but by activation of the JAK/STAT3 pathway by signal transduction proteins like NF-κB (51) which our group has shown is transcriptionally regulated by REV-ERB-α in microglia (3), and which was also a hit in our bioinformatic analysis described herein. Further studies are necessary to determine the cell-type specificity of REV-ERB-α/-β regulation of Stat3 expression in the brain given its tissue-specific cistrome (52), as well as alternative upstream drivers that could participate in the phenotypes observed in the brain following REV-ERB-α/- β deletion.

Our study has several limitations. We have not directly addressed the REV-ERB-β-specific transcriptome in the brain with REV-ERB-β knockouts. Further studies are needed to determine if the effects observed in our study are the result of specific reduction of both REV-ERB-α/-β or simply a reduction in REV-ERB total dosage levels to a threshold where the remaining REV-ERBs expression is insufficient to maintain basal transcriptional repression. Cre lines also have varying degrees of expression which influence the degree of gene deletion and transcriptional changes. Furthermore, the Aldh1l1-Cre lines do express highly in the liver, which might also impact brain phenotypes. Despite this weakness, this line expresses much more broadly across astrocytes and is quite specific in the brain compared to other lines. Finally, because the CAG-α/βKO mice in Fig. 1 were harvested at 8 to 9 mo (6 to 7 mo after tamoxifen), while all other mice in the study were harvested at 4 mo (2 mo after tamoxifen), gene expression cannot be directly compared between CAG-α/βKO mice in Fig. 1 and Aldh1l1-α/βKO mice in Fig. 3. However, gliosis was quantified in 4mo CAG-DKO mice in Fig. 2 which can be compared with Fig. 3.

In summary, this study shows that REV-ERB-α/-β play partially redundant roles in the CNS and that deletion of both REV-ERB-α and REV-ERB-β unmasks a robust transcriptomic signature that is not seen by targeting REV-ERB-α alone. Efforts to utilize REV-ERB-α/-β should consider both isoforms when targeting putative pathways regulated by REV-ERB-α/-β. Additionally, we have shown that REV-ERB-α/-β coordinately regulate astrocyte reactivity in a cell-autonomous manner that is associated with increased Stat3 expression and enhances proteostasis, resulting in protection against alpha-synucleinopathy. Our findings illuminate REV-ERBs as important regulators of astrocyte reactivity and suggest that REV-ERB inhibition may be a plausible therapeutic target for neurodegenerative diseases.

Methods

Mice.

All mouse experiments were conducted according to protocols approved by the Washington University Institutional Animal Care and Use Committee (IACUC) and were under the supervision of the Division of Comparative Medicine (DCM). REV-ERB-α floxed (Nr1d1^fl/fl) and REV-ERB-β floxed (Nr1d2^fl/fl) mice on C57Bl/6 background have been previously described (53, 54). CAG-Cre^ERT2+ and Aldh1l1-Cre^ERT2+ mice were obtained from Jackson Laboratory (Bar Harbor, ME). All mice were housed in constant conditions in 12:12 h light–dark cycle with ad libitum food and water. Briefly, CAG-Cre⁺ males were crossed with Nr1d1^fl/fl or Nr1d1^fl/fl; Nr1d2^fl/fl females to generate CAG-Cre^ERT2+; Nr1d1^fl/fl or CAG-Cre^ERT2+; Nr1d1^fl/fl; Nr1d2^fl/fl lines. Similar breeding scheme was used to generate Aldh1l1-Cre^ERT2+; Nr1d1^fl/fl; Nr1d2^fl/fl lines. Littermates from multiple litters were used in analysis. Mice were treated with tamoxifen (Sigma, 5648, solubilized in corn oil, 4 mg/day/mouse × 5 d) via oral gavage at 2 to 3 mo old to induce deletion of floxed genes. Mice were harvested 2 to 3 mo post tamoxifen administration for the younger cohort and 6 mo post tamoxifen for the older cohort, unless otherwise stated. All mouse cohorts were mixed sex, with at least 2 mice of each sex in each genotype. While not powered to detect sex differences, no obvious differences were noted.

Immunohistochemistry and Imaging.

Mice were deeply anesthetized via intraperitoneal pentobarbital (150 mg/kg), then perfused with ice-cold Dulbecco’s modified PBS (DPBS) containing 0.3% heparin (v/v). For non-PFF injected mice, one hemisphere was postfixed in 4% paraformaldehyde (PFA) for 24 h (4 °C), then cryoprotected with 30% sucrose in PBS (w/v, 4 °C) until sectioning. The other hemisphere was dissected into cortical, hippocampal, and subcortical regions and were flash frozen and stored at −80 °C until RNA or protein workup. For PFF-injected mice, the entire brain was collected and postfixed in PFA for 24 h and then cryopreserved in sucrose as previously described. Brains were sectioned on a freezing sliding microtome (Leica) in 40-micron serial coronal sections, and sections were stored in cryoprotectant (30% ethylene glycol, 15% sucrose, 15% phosphate buffer in ddH20) at −20 °C until used for staining. A minimum of 2 sections per mouse were used as technical replicates in all experiments. Free-floating sections were washed 3× in TBS and blocked in 3% goat or donkey serum in TBS with 0.25% Triton X-100 (TBSX) for 1 h at room temperature (RT). Sections then incubated with the following primary antibodies overnight at 4 °C in TBSX with 1% goat or donkey serum: GFAP (rabbit, Dako/Agilent Z0334, 1:3,000), IBA1 (goat, Abcam ab5076), IBA1 (rabbit, Wako, 019-19741, 1:500), CD68 (rat, BioRad MCA1957, 1:500), conjugated GFAP conjugated to AlexaFluor-647 (mouse, Cell Signaling Technologies, 3657S, 1:800), biotinylated phospho-alpha-synuclein (mouse, Biolegend 825704, 1:1,000).The following day, sections were washed 3× in TBS then incubated in corresponding goat or donkey serum with fluorescent secondary antibodies for 1 h at room temperature. Sections then mounted on slides and coverslipped with Prolong Gold (Invitrogen P36930).

Fluorescent sections were imaged with epifluorescence on the Keyence BZX-810 microscope. Confocal imaging was performed on a Nikon AXR at 40× magnification. Light intensity, laser power, and camera exposure time were optimized for each imaging cohort. All images analyzed together were captured at the same time with the same imaging settings.

Epifluorescent images were analyzed using FIJI (FIJI is Just ImageJ). Briefly, TIFF files were converted to 8-bit grayscale images. Multiple images of varying staining intensity were used to optimize the brightness, background subtraction, and thresholding values that were used for the entire imaging cohort. Fluorescent intensity of images was thresholded to determine the percent area of the region of interest using the Analyze Particles function in FIJI. 3D image reconstructions and colocalizations were generated from confocal z-stacks using Imaris (Oxford Instruments). Volume statistics abstraction was done using Imaris. For hippocampal images, the entire anterior hippocampus was imaged in multiple stitched images and quantification included all areas.

RNA Extraction and Gene Expression Analysis.

Cortical brain tissue was homogenized in 500 of TRIzol (ThermoFisher) with RNase-free zirconium oxide beads (Next Advance) in a bullet blender (Next Advance) for 3 min. The entire cortex was removed and homogenized. Chloroform was added to TRIzol samples in a 1 chloroform: 6 total volume ratio, and samples were thoroughly mixed and centrifuged at 12,000×g for 10 min at 4 °C. RNA-containing aqueous supernatant was collected, and then RNA was extracted and purified using the PureLink RNA Mini Kit (ThermoFisher) according to the manufacturer’s protocol. Sample RNA concentration and purity was measured using a Nanodrop spectrophotometer, cDNA was made using a high-capacity RNA-cDNA reverse transcription kit (Applied Biosystems) with 1,000 ng RNA input per 20 μL reaction. Real-time qPCR (RT-qPCR) was performed with TaqMan primers (Applied Biosystems) and PCR Master Mix buffer (Applied Biosystems) on StepOnePlus thermocyclers (Applied Biosystems). Original relative mRNA expression was normalized to β-actin (Actb) mRNA expression or analysis. To perform a larger transcriptomic profiling, microfluidic qPCR assay was performed by the Washington University Genome Technology Access Center (GTAC) using a Fluidigm Biomark HD system.

RNA sequencing was performed by GTAC using the isolated cortical RNA as described previously (12, 46). DESeq2 package in R was utilized to obtain DEG lists after accounting for multiple comparison with normalized log₂ fold change and adjusted P values (q value) to account for multiple comparisons. Volcano plots were generated using the EnhancedVolacno R package and heatmaps were generated in GraphPad Prism 10. KEGG pathway plots were generated with EnrichKEGG package in R. All RNAseq raw data are publicly available on the NIH GEO repository under GEO accession numbers GSE293033 and GSE293026.

Western Blots.

Hippocampal tissue was homogenized in RIPA buffer with protease/phosphatase inhibitor (ThermoFisher). Protein concentration was determined with Pierce BCA Protein Assay Kit (ThermoFisher) and 5 μg of total protein was loaded and run on a 4 to 12% Bis-Tris gradient gel. Protein was then transferred to a polyvinylidene difluoride membrane (PVDF), and membranes were blocked in 5% bovine serum albumin (BSA) in TBS with 0.05% Tween (TBS-T) for 1 h at RT and then washed in TBS-T. Membranes were then incubated in TBS-T with 1% BSA overnight at 4 °C with the primary antibodies REV-ERB-α (Invitrogen, PA5-29865, 1:1,000), REV-ERB-β (Santa Cruz, sc-398252, 1:1,000), STAT3 (Cell Signaling, 8768, 1:1,000), phosphorylated-STAT3 (Cell Signaling, 9145, 1:2,000), β-actin (Cell Signaling, 4970, 1:8,000). The following day, membranes were then washed three times in TBS-T and then incubated in secondary antibody (Cell Signaling, anti-rabbit IgG, HRP-linked, 7074, 1:1,000). Blots then developed with Lumigen ECL Ultra (ThermoFisher, TMA-100) and imaged on BioRad ChemiDoc Imaging System. Densitometric analysis of band signal intensity was performed in FIJI with all quantified bands normalized to β-actin intensity.

Primary Mouse Astrocyte Culture.

Primary mouse astrocyte-enriched cultures were prepared as previously described (12). Briefly, P1-P3 CD1 mouse pup brains were removed in ice-cold DMEM, the meninges carefully peeled away under a microscope, and the cortices dissected apart from the hippocampus and subcortical structures. Cortices were then enzymatically digested in 0.025% trypsin solution (Gibco) at 37 °C for 15 min, then triturated into a single cell suspension using a pipet and then seeded in a poly-D-lysine (PDL)-coated T75 flask. Cells grown in DMEM with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (“astrocyte media”). Once cells reached 90 to 100% confluency, flasks were shaken at 225 RPM at 37 °C for 2 to 3 h to remove overlaying microglia and enrich for astrocytes. Cells were then plated in PDL-coated plates and allowed to grow to confluency in the plates prior to start of experiments.

siRNA Transfections.

Upon confluency in the well plates, primary mouse astrocyte-enriched cultures were transfected with siRNA using Lipofectamine RNAiMAX (Life Technologies) in OptiMEM (Life Technologies) per the manufacturer’s guidelines. Astrocytes incubated in transfection media for 24 h and then transfection media were changed back to normal astrocyte media (DMEM plus 10% FBS and 1% penicillin/streptomycin). siRNAs targeting mouse Nr1d1 (L-051721-00-0005), Nr1d2 (L-059128-01-0005), and scrambled nontargeting control (SCR, D-001810-10-05) were obtained from Horizon Discovery/Dharmacon (Lafayette, CO). ON-TARGETplus SmartPool siRNA was used, which pools together four separate siRNAs targeting the gene of interest to ensure sufficient gene knockdown. An siRNA:RNAiMAX ratio of 1:1.25 was used, and 40 pmol of siRNA was added to each well 24 h before replacing with normal astrocyte media.

Phagocytosis Assay.

Primary astrocytes were seeded at 60,000 cells per well in a 24-well plate and allowed to grow to confluency. Once confluent, siRNAs were introduced and removed the following day. Three days after siRNA removal, PFFs conjugated to pHrodo Deep Red and AlexaFluor488 were added at a concentration of 4.0 µg/mL in astrocyte media for 4 h. Conjugated PFFs were graciously generated and given by the Davis Lab (12). For the lysosomal inhibition experiment, chloroquine (20 µM in astrocyte media) was added 1 h before the addition of PFFs and remained present during coincubation with PFFs. Following PFF removal, cells were rinsed once with DPBS, trypsinized, washed with serum-free DMEM, resuspended in flow buffer (DPBS+1% FBS+1 mM EDTA), and analyzed using a Beckman Coulter CytoFLEX S Flow Analyzer. Data analysis was performed using FlowJo software (BD Biosciences).

Alpha-Synuclein PFF Stereotactic Injection.

Purified mouse alpha-synuclein (αSyn) PFF were generously prepared and given by the Davis Lab as previously described (12). Before injections, PFF aliquots were sonicated for 8 min in a water bath sonicator (Qsonica). Mice were anesthetized with isoflurane, placed on a stereotactic apparatus (Kopf), and stereotactically injected with 2 μL (approximately 10 μg of PFFs) into the right striatum using a Hamilton microsyringe (#1702) attached to a motorized injector (Stoelting). The coordinates used were 0.2 mm anterior and 2.0 mm lateral to bregma, and 3.2 mm ventral to the skull surface. The infusion was carried out over a 5-min interval, and the needle was left in place for 4 min to allow the liquid to diffuse after the injection before the needle was slowly withdrawn. The scalp was then sutured and closed with VetBond. Buprenorphine sustained release (1 mg/kg) was given i.m. for long-term analgesia.

Computational Prediction of Putative Transcription Factors.

Lists of DEGs were used as input to predict putative transcription factors (TF) regulating the set of query gene sets. Prediction were performed using four computational tools: MORA (29), HOMER (55), BART2 (56), and LISA2 (57). Due to the unique format of the output file of HOMER which lacks shared identifier with the other tools, EnsembleTFpredictor (29) was employed to integrate prediction results from all four tools. Predicted TFs were ranked based on the number of tools that identified them as statistically significant.

Statistics and Graphs.

All experimental schematics were created using BioRender. All graphs in figures depict the group mean ± SEM and N represents the individual number of biological replicate animals unless otherwise stated. For cell culture experiments, representative graph and analysis of one experiment is shown and graphs show group of average of technical well replicates as stated in figure legends. For datasets with one independent variable, unpaired 2-tail t test with Welch’s correction for unequal SD or one-way ANOVA with multiple comparisons were performed. If variances of the individual groups were significantly different by F-test, then the nonparametric Mann–Whitney U-test was performed as noted. For datasets with two independent variables, two-way ANOVA with Tukey’s multiple comparison was performed. Outliers were identified by Grubb’s test and were excluded from analysis. Statistical tests were performed using GraphPad Prism software (Version 10.4.1). All graphs were created using GraphPad Prism. P values greater than P > 0.1 were noted as not significant and were not specifically denoted. P values between 0.05 and 0.1 were noted in the figure. P values < 0.05 were denoted statistically significant and noted with asterisk, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Diagrams in Figs. 1, 3, and 5 were created in BioRender. Musiek, E. (2025) https://BioRender.com/hm4sm29.

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Data, Materials, and Software Availability

RNAseq data have been deposited in NIH GEO (GSE293033 and GSE293026) (20, 21).