Neuronal deletion of the circadian clock gene Bmal1 induces cell-autonomous dopaminergic neurodegeneration

Circadian clocks regulate many aspects of cellular function, including cellular metabolism, redox homeostasis, and inflammation (1, 2). On a molecular level, the core circadian clock consists of a transcriptional-translational feedback loop that tunes gene expression and cellular function to a 24-hour rhythm. This core circadian clock consists of the basic helix-loop-helix transcription factor BMAL1 (also referred to as ARNTL), which heterodimerizes with CLOCK or NPAS2 to drive tissue-specific transcriptional programs. BMAL1 drives expression of several negative-feedback repressors, including PER, CRY, and REV-ERB proteins, which in turn inhibit BMAL1-mediated transcription and are tuned through posttranscriptional regulation to a 24-hour period (3). BMAL1 is required for circadian oscillations in transcription, cellular function, and behavior, and Bmal1 deletion renders mice arrhythmic at both cellular and behavioral levels (4). While binding of BMAL1 to DNA is highly rhythmic, BMAL1 also regulates expression of many nonrhythmic transcripts and is, thus, thought to exert additional “noncircadian” functions (5, 6). Our group has previously shown that Bmal1 deletion causes astrogliosis, oxidative stress, and synapse loss in mouse cerebral cortex (6, 7). However, overt neurodegeneration was not observed in Bmal1-KO mice in vivo in regions such as the hippocampus, cortex, and striatum (7). Furthermore, Bmal1 deletion appears to have protective effects in some models of neurological disease, such as tau and α-synuclein (αSyn) aggregation, stroke, and spinal cord injury (8–10). Thus, the role of Bmal1 in brain health and neurodegeneration is not entirely understood.

Parkinson disease (PD) is a common neurodegenerative disease characterized clinically by progressive tremor, rigidity, bradykinesia, postural instability, autonomic dysfunction, and neuropsychiatric symptoms. Key pathological hallmarks of PD include degeneration of dopaminergic neurons in the substantia nigra pars compacta (SNpc) and intraneuronal accumulation of aggregated αSyn in Lewy bodies and Lewy neurites. A considerable amount of literature describes circadian disruption in PD, including fragmentation of daily behavioral rhythms and blunting of rhythms of Bmal1 and other core circadian clock gene expression in peripheral blood mononuclear cells (PBMCs) from patients with PD (11–15). In mice, both light-induced circadian disruption (16) and germline Bmal1 deletion have been shown to exacerbate dopaminergic neurodegeneration caused by the mitochondrial toxin 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (17). While a previous study implicated microglial dysfunction as critical to this phenotype, germline Bmal1-KO mice exhibit a variety of developmental phenotypes that could contribute, and a comprehensive cell type–specific analysis has not been conducted in vivo (17). Here, we examine the effects of postnatal and cell type–specific Bmal1 deletion on viability of dopaminergic neurons in the SNpc and test the effect of Bmal1 deletion on αSyn pathology and neurodegeneration in a mouse model. Our data indicate that Bmal1 regulates specific cell-autonomous transcriptional programs in dopaminergic neurons that are critical for their survival, suggesting that preserving Bmal1 expression may be neuroprotective in PD.

We first examined the number of midbrain dopaminergic neurons in mice following tamoxifen-inducible, postnatal global Bmal1 deletion (CAG-Cre^ERT2;Bmal1^fl/fl). We and others have previously characterized this mouse line and have shown that postnatal tamoxifen administration causes widespread Bmal1 deletion in the brain as well as behavioral arrhythmicity, while avoiding multiple developmental complications noted in germline Bmal1-KO mice (18, 19). We performed unbiased stereology of tyrosine hydroxylase⁺ (TH⁺) neurons in the SNpc in Cre^– (control) and Cre⁺ (global Bmal1-KO) littermates, all treated with tamoxifen at 2–4 months old. One month after tamoxifen treatment, we observed a loss of BMAL1 immunoreactivity in the nuclei of TH⁺ neurons in the SNpc of Cre⁺ mice (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.162771DS1↗). We observed a 40% decrease in the number of TH⁺ neurons in Cre⁺ mice compared with Cre^– mice (Figure 1, A–C). TH immunoreactivity in the striatum was also decreased by 41% in Cre⁺ mice compared with Cre^– mice (Figure 1, B and D). We also observed a reduction in number of NeuN⁺ neurons in the SNpc in CAG-Cre⁺ mice, and this was similar in magnitude as compared with the loss of TH⁺ cells, suggesting that neurodegeneration — rather than simply a reduction in TH protein expression — accounted for the apparent loss of TH⁺ neurons (Figure 1E). We also counted TH⁺ neurons in the ventral tegmental area (VTA), an adjacent brain region containing dopaminergic neurons but one that does not exhibit the same selective vulnerability as the SNpc in PD. TH⁺ neuron numbers were also reduced in the VTA (Supplemental Figure 1, B and C), suggesting a broad sensitivity of TH⁺ neurons to global deletion of Bmal1. However, NeuN⁺ neuronal nuclei were grossly intact in the hippocampus of Cre⁺ global KO mice compared with Cre^– mice, suggesting that hippocampal neurons do not degenerate in similar numbers as TH⁺ neurons, though detailed stereology was not performed (Supplemental Figure 1D). A previous study showed that REV-ERBα, a direct transcriptional target of BMAL1 that is suppressed in Bmal1-KO brain (7), suppresses Th gene expression in the midbrain (20). Thus, we would not expect that Bmal1 deletion would directly suppress Th expression; rather, increased expression of Th gene might be expected. Accordingly, there was no change in Th mRNA levels in Nestin-Cre;Bmal1^fl/fl mouse cortex (Supplemental Figure 1E). Thus, it is unlikely that Bmal1 deletion simply causes downregulation of the Th gene. Of note, neither CAG-Cre^ERT2+;Bmal1^fl/fl mice, which were not treated with tamoxifen, nor tamoxifen-treated CAG-Cre^ERT2+ mice, in which Bmal1 is not floxed, showed any loss of TH⁺ neurons in the SNpc, demonstrating that Cre-related toxicity did not explain our findings (Supplemental Figure 2).

Global inducible Bmal1-KO mice are behaviorally arrhythmic under constant lighting conditions and have Bmal1 deletion in all tissues (18, 19). In order to determine if disrupted behavioral rhythms or peripheral Bmal1 deletion contributed to TH⁺ neuron loss, we examined the SNpc of 9- to 11-month-old Nestin-Cre⁺;Bmal1^fl/fl mice, which have deletion of Bmal1 in neurons, astrocytes, and other cells throughout the brain but have intact Bmal1 expression in peripheral organs as well as intact behavioral rhythms due to limited Cre-mediated deletion of Bmal1 in the suprachiasmatic nucleus (SCN) (7, 18). Nestin-Bmal1–KO mice also exhibited a similar degree of loss of TH⁺ neurons in the SNpc and loss of TH staining in the striatum (Figure 1, F–I), demonstrating that degeneration is likely due to loss of Bmal1 expression in neurons or astrocytes and not due to disrupted whole-animal behavioral circadian rhythms or peripheral organ dysfunction. Notably, Nestin-Bmal1–KO mice also retain microglial Bmal1 expression, indicating that microglial Bmal1 deletion is not required for the TH neurodegeneration phenotype.

In order to further investigate the role of whole animal circadian rhythm disruption in dopaminergic neurodegeneration, we subjected 3-month-old WT mice to circadian desynchrony via 12 weeks of exposure to a 10-hour/10-hour light/dark (L:D) cycle (T20 cycle) (21). Control mice were kept on standard 12-hour/12-hour L:D lighting. This intervention causes circadian desynchrony between the light cycle and the internal clock, and it has been shown to cause alterations in neuronal morphology and cognitive impairment in rodents (22). We observed disruption of behavioral circadian rhythms in mice subjected to the T20 cycle (Figure 2A), but this had no effect on TH⁺ neuron counts in the SNpc (Figure 2, B and C). This result suggests that disruption of behavioral circadian rhythms via altered lighting schedule does not phenocopy the effect of Bmal1 deletion on TH neuron viability.

We next counted SNpc TH⁺ neurons in mice with neuron–specific Bmal1 deletion. We used a particular line of BAC-transgenic CaMK2a-Cre;Bmal1^fl/fl mice, previously shown to have widespread, pan-neuronal Bmal1 deletion and behavioral rhythm disturbances (6, 22). Both 4-month-old and 10-month-old CaMK2a-iCre;Bmal1^fl/fl mice exhibited a reduction in number of SNpc TH⁺ neurons as compared with Cre^– littermate controls (Figure 3, A and B), demonstrating a neuronal cell–autonomous effect of Bmal1 deletion on TH⁺ cell survival. There was no main effect of age (4-month-old versus 10-month-old) on neuronal loss in CaMK2a-iCre;Bmal1^fl/fl mice and no interaction between genotype and age, suggesting that degeneration due to Bmal1 deletion occurs early and then plateaus.

We have previously shown that mice with astrocyte-specific Bmal1 deletion (Aldh1l1-Cre^ERT2;Bmal1^fl/fl) develop spontaneous astrogliosis and astroglial transcriptional alterations (6). To evaluate a role for astrocyte Bmal1 in dopaminergic cell loss, we next examined TH⁺ neurons in the SNpc in Aldh1l1-Cre^ERT2;Bmal1^fl/fl mice. One month after tamoxifen treatment, 4- to 5-month-old astrocyte Bmal1-KO mice showed no change in TH⁺ neuron counts in the SNpc compared with Cre^– littermate controls (Figure 3, C and D). Thus, Bmal1 deletion in astrocytes does not account for the TH⁺ neuron loss phenotype observed in Nestin-Bmal1–KO and CAG-Bmal1–KO mice.

In order to examine possible transcriptional mechanisms by which Bmal1 deletion might influence neuronal survival, we performed bulk RNA-Seq on cerebral cortex tissue from 4-month-old pan-neuronal Bmal1-KO mice (Camk2a-Bmal1–KO [Bmal1^fl/fl]) and Cre^– littermate controls (Figure 4A). The known BMAL1 transcriptional targets Nr1d1 and Dbp were strongly downregulated in Cre⁺ mouse cortex, indicating loss of BMAL1 activity (Figure 4A). Utilizing differential expression analysis, we identified 634 dysregulated genes, using an unadjusted P < 0.05 cutoff. Notably, the number of differentially expressed genes (DEGs) was substantially lower than that seen in global Bmal1-KO mice, and astrocyte-specific BMAL1 target transcripts such as Chi3l1 — which is differentially expressed in global KOs (23) — were unchanged in the Camk2a-Bmal1–KO mice (Figure 4, A and B). Thus, the transcriptome appears quite specific for neuronal Bmal1 deletion. KEGG pathway gene set enrichment analysis (GSEA) was then performed on the DEGs in this data set, and it identified PD and oxidative phosphorylation as 2 of the top 5 pathways (Figure 4C). Thus, Bmal1 deletion in neurons causes dysregulation of transcriptional pathways involved in oxidative phosphorylation and PD in cortex.

In order to determine if this gene expression profile in Bmal1-KO tissue is relevant to the midbrain, where TH⁺ neuronal loss is observed, we dissected midbrain samples from CAG-Cre^ERT2;Bmal1^fl/fl mice and Cre^– littermate controls 2 months after tamoxifen treatment (5 months old) and performed transcriptomics analysis via bulk RNA-Seq. Unfortunately, the sequencing from 2 samples did not meet quality-control standards, so we ultimately examined 3 Cre^– and 2 Cre⁺ samples. However, we performed exploratory analyses on these data to determine if they showed similar trends as the data from Bmal1-KO cortex described above. We examined several known DEGs from previous studies of Bmal1-KO mouse brain and observed the expected trends in expression of Nr1d1, Dbp, and Chi3l1 as would be seen with Bmal1 deletion (7, 23) (Supplemental Figure 3A). Exploratory differential gene expression analysis identified 2,444 DEGs that met an unadjusted P < 0.05 and 371 that reached a multiple-comparisons adjusted P < 0.05 (Figure 4D). We first performed KEGG pathway analysis on this preliminary list of 371 DEGs and identified several specific pathways including oxidative phosphorylation, PD, Huntington’s disease, Prion disease, neurodegeneration, ROS, and Alzheimer’s disease (Figure 4E). We also examined the 2,444 gene group with unadjusted P < 0.05 with KEGG pathway analysis. Genes that were upregulated in Cre⁺ mice again showed enrichment for transcripts involved in ribosome function, PD, oxidative phosphorylation, and Huntington’s disease (Supplemental Figure 3B). Pathways that were downregulated genes in Cre⁺ were enriched for transcripts related to cGMP/PKG signaling, glutamatergic synapses, and circadian function (Supplemental Figure 3C). Expression of specific transcripts in the PD KEGG pathway are shown in Supplemental Figure 4D. Thus, while not statistically rigorous, our exploratory transcriptomics analysis of global Bmal1-KO midbrain suggests that similar pathways are altered as in our analysis of neuron-specific Bmal1-KO cortex and that oxidative phosphorylation and PD are prominent.

Finally, we sought to determine if deletion of Bmal1 specifically in TH⁺ neurons was sufficient to induce cell-autonomous degeneration. Accordingly, we generated TH-Cre;Bmal1^fl/fl mice to drive Bmal1 deletion exclusively in TH⁺ cells. We confirmed by confocal microscopy that BMAL1 protein expression was absent in some but not all TH⁺ cell nuclei (Figure 5, A–C), indicating deletion within a subset of TH⁺ neurons of the SNpc (approximately 25% of cells, on average). The actual deletion of BMAL1 may occur in a higher percentage of cells, since the best antibody we could find gives a slight residual nuclear signal even in germline Bmal1-KO mice. However, because of this incomplete deletion, we referred to this as Bmal1 knockdown (KD), rather than KO. Unbiased stereology revealed that TH-Bmal1 KD mice showed a 23% reduction of TH⁺ neurons in the SNpc, as compared with Cre^– controls, demonstrating that Bmal1 exerts cell-autonomous effects on dopaminergic neuron survival in vivo (Figure 5D). However, TH immunoreactivity in the striatum was not changed in Cre⁺ mice (Figure 5E). Upon further analysis, we also found a decrease in Iba1⁺ microglia branching within the SNpc in Cre⁺ mice (Figure 5F), indicating a transition to a more amoeboid morphology associated with microglial activation, around the site of TH⁺ neurodegeneration. Microgliosis was not observed in the hippocampus of Cre⁺ mice (Supplemental Figure 4), suggesting a region-specific effect. Notably, we observed no changes in circadian activity rhythms in TH-Cre⁺;Bmal1^fl/fl mice under L:D or constant dark (D:D) conditions (Figure 5, G and H), further demonstrating that the effects are not due to changes in circadian behavioral rhythms.

A previous study showed that germline Bmal1 deletion can exacerbate dopaminergic neuron loss caused by the mitochondrial complex I inhibitor MPTP (17). To test this relationship in another model system relevant to PD, we examined TH⁺ neuron survival in CAG-Bmal1–KO mice following intrastriatal injection of αSyn preformed fibrils (PFFs). In this model, exogenous αSyn PFFs are injected into the striatum and are likely transported in a retrograde manner to the SNpc, inducing misfolding of endogenous αSyn and causing dopaminergic neurodegeneration (24, 25). We treated mice with tamoxifen at 2 months old, injected αSyn PFFs into the right striatum at 3 months old, and euthanized the mice at 6 months old. We again observed a loss of TH⁺ neurons in the Cre⁺ mice on the contralateral side, demonstrating basal TH neuron loss as a result of Bmal1 deletion, as seen in Figure 1. αSyn PFF injection caused a similar proportional loss of TH⁺ neurons in the ipsilateral SNpc of Cre^– control mice compared with Cre⁺ mice (Figure 6, A and B). Thus, Bmal1 KO does not seem to exacerbate αSyn PFF–induced TH neuron loss, though Cre⁺ mice ended up with fewer TH⁺ neurons because they had fewer neurons at baseline. We have previously shown that Bmal1 deletion can prevent αSyn seeding and spreading (8) but observed no difference in pSer129- αSyn IHC staining in the SNpc between genotypes (Figure 6C). This is likely because PFFs enter TH⁺ nerve terminals in the striatum directly after injection and cause SNpc injury, without the need for seeding or spreading. Striatal TH immunoreactivity was decreased in Cre⁺ mice but was not afftected by PFF injection (Figure 6, D and E). We also tested locomotor activity in tamoxifen-treated CAG-Cre^ERT2+;Bmal1^fl/fl and Cre^– controls 80 days after PBS or PFF injection using the pole-descending test, which is sensitive to dopaminergic cell loss (Supplemental Figure 5) (26). While we observed an effect of Bmal1 genotype on this behavior, there was no effect of PFF injection and no interaction between genotype and PFF treatment. CAG-Bmal1–KO mice surprisingly performed better on this motor task compared with Cre^– controls, suggesting that their behavioral changes were not driven by loss of dopaminergic neurons. The lack of behavioral effect after PFF injection is presumably because the degree of TH⁺ neuron loss in this particular cohort was not large enough to cause major motor symptoms, as ~80% neuron loss has been suggested as necessary to cause motor deficits (26). It is also notable that unilateral injections, as done here, are less effective than bilateral injections at producing behavioral deficits (27).

A previous study suggests that loss of TH⁺ neurons in global Bmal1-KO mice in response to MPTP may be mediated by microglia (17). To determine if any pathological phenotype is mediated by loss of microglial Bmal1 expression, we employed the same α-syn PFF model using microglia-specific Bmal1-KO mice (Cx3cr1-Cre^ERT2;Bmal1^fl/fl). First, no loss of neurons was observed on the contralateral side when comparing Cre^– control and Cre⁺ microglial Bmal1-KO mice, suggesting that spontaneous TH⁺ neuronal loss is not mediated by Bmal1-deficient microglia (Supplemental Figure 6). Second, microglia-specific KO mice showed no difference in TH⁺ cell number or phosphosynuclein staining between genotypes on the ipsilateral side of the injection (Supplemental Figure 6), further supporting a cell-autonomous effect of neuronal Bmal1 on dopaminergic neurodegeneration.

Our data suggest that the circadian clock transcription factor BMAL1 plays a cell-autonomous role in regulating dopaminergic neuron viability in the SNpc. This effect is not replicated by light-mediated disruption of behavioral circadian rhythms or by glia-specific Bmal1 deletion. Neuronal BMAL1 regulates transcriptional pathways related to PD and oxidative phosphorylation, and neuronal Bmal1 deletion leads to spontaneous death of a portion of TH⁺ neurons. Our findings reveal BMAL1 as a contributor to dopaminergic neuron survival and suggest that environmental or disease states that reduce BMAL1 expression might trigger or exacerbate neurodegeneration and parkinsonian phenotypes.

Previous reports have demonstrated that both altered 24-hour light cycle (16) and global germline Bmal1 deletion (17) can exacerbate TH⁺ neuronal loss in the SNpc in response to the mitochondrial toxin MPTP. Our findings expand upon these observations by using multiple cell- and tissue-specific Bmal1–KO mice to show spontaneous TH neuron loss following neuronal Bmal1 deletion. Loss of microglial Bmal1 was implicated as the mechanism of degeneration in one of these previous studies, based on cell culture experiments (17). However, our data showing TH⁺ neuron loss in Nestin-Bmal1–KO mice (which spares microglial Bmal1 expression), as well as our observation that microglia-specific Bmal1 deletion did not alter either basal or αSyn PFF–induced TH⁺ neuronal loss, indicates that microglial Bmal1 does not drive the neurodegeneration phenotype, at least in our models. Astrocyte-specific Bmal1 deletion can cause astrogliosis (6), but deletion of Bmal1 in astrocytes also had no effect on dopaminergic neuronal survival in our experiments. Moreover, we did not observe loss of TH⁺ neurons after 3 months of light-induced circadian desynchrony (T20 cycle). Because numerous light-based circadian desynchrony models exist, each with potentially unique effects, further study is needed to understand the effect of altered light cycles on dopaminergic neurodegeneration. However, our data point strongly to neuronal loss of Bmal1 as being fundamental to this TH neuronal death phenotype, independent of changes in behavioral circadian rhythms. While this effect of Bmal1 deletion may be independent of its role in circadian rhythms, we cannot exclude the possibility that disruption of cell-intrinsic circadian rhythms in TH neurons via Bmal1 deletion contributes to cell death.

Our findings build on previous studies from our laboratory, suggesting that BMAL1 plays a role in neuronal redox homeostasis. Global Bmal1-KO mouse cortex shows increased lipid peroxidation and impaired circadian expression of redox defense transcripts such as Nqo1 and Aldh2 (7) as well as diminished expression of several glutathione S-transferase transcripts (6). Our RNA-Seq analysis shows that BMAL1 regulates transcripts related to PD pathogenesis and oxidative phosphorylation in midbrain tissue from global Bmal1-KO mice and in cortical tissue from neuron-specific Bmal1-KO mice. Dopaminergic neurons of the SNpc are uniquely sensitive to oxidative stress and mitochondrial dysfunction, and this may explain why neuronal loss was noted in the SNpc but not in the hippocampus or other regions in Bmal1-KO mice. Accordingly, we previously demonstrated that, while Bmal1-deficient cortical neuron cultures were sensitized to hydrogen peroxide–induced cell death, global Bmal1 deletion did not affect basal neuronal survival in the cortex and striatum in vivo. However, hemizygous deletion of Bmal1 sensitized striatal neurons to death induced by the mitochondrial toxin 3-nitropropionic acid (7). Thus, Bmal1 deletion may render all neurons more vulnerable to oxidative stress, which could specifically manifest as TH neuron loss.

Our findings emphasize that the effects of BMAL1 in neurological disease models are highly cell type and context dependent. Global Bmal1 deletion has previously been shown to cause brain oxidative stress and synapse loss (7), exacerbate MPTP-induced dopaminergic neuron loss(17), accelerate amyloid plaque formation (18), and lower seizure threshold (28, 29) — all effects that should promote neurodegeneration. Conversely, previous studies show that global Bmal1 deletion can mitigate stroke severity (9) and spinal cord injury(10). We have previously shown that Bmal1 deletion specifically in astrocytes induces astrogliosis (6) and can prevent both tau and αSyn aggregation in vivo, in part through regulation of the chaperone protein BAG3 or through changes in lysosomal function (8, 30). Similarly, other groups have demonstrated that deletion of Bmal1 in microglia appears to reduce inflammatory activation and reduce damage caused by ischemic stroke (31). Deletion of Bmal1 in either lymphocytes or myeloid cells has also been shown to also reduce disease severity in the experimental autoimmune encephalomyelitis (EAE) model of multiple sclerosis (32, 33). Our data in this manuscript showing loss of TH neurons following global or neuronal Bmal1 deletion (Figure 1, Figure 2, and Figure 4) are consistent with our previous study showing that neuronal Bmal1 deletion sensitizes neurons to oxidative stress and death (7), while Bmal1 deletion in glia may have more variable (and often protective) effects (8). Thus, BMAL1 appears to be important for neuronal survival in general. Further study into the cell- and model-dependent effects of Bmal1 and other clock genes is needed to fully understand how core clock disruption may influence neurodegenerative diseases.

Sleep and circadian disturbances are common in synucleinopathies including PD. Several human studies show alterations in the expression of clock gene transcripts in PBMCs from patients with PD. Bmal1 transcript levels in PBMCs were suppressed in patients with PD in 2 studies (14, 15), while the rhythmic Bmal1 expression was blunted in patients with PD in another (13). While it is unknown how peripheral blood and SNpc Bmal1 expression are correlated in patients with PD, our studies suggest that environmental factors and disease processes that affect circadian gene expression in the brain, in particular those that lead to suppressed Bmal1 expression, could put dopaminergic neurons at risk of death. Conversely, therapies that augment Bmal1 expression might be expected to protect TH⁺ neurons, and this possibility warrants future investigation.

In summary, the circadian clock gene Bmal1 exerts cell-autonomous effects on dopaminergic neuronal survival. Neuronal Bmal1 regulates transcripts related to oxidative phosphorylation and PD, and loss of neuronal Bmal1 leads to death of TH⁺ neurons in the SNpc. These effects are not phenocopied by light-induced circadian desynchrony or by glial Bmal1 deletion. Our results provide insights into the relationship between the neuronal circadian clock and PD pathogenesis.

MKK and PWS share first-author position. MKK was listed first because he generated a larger number of the figures, and PWS agreed to this arrangement. AAD and ESM share corresponding author position, as AAD was responsible for overseeing TH stereology and synuclein models, while ESM managed all aspects of mouse model generation and RNA-Seq, with both authors sharing conceptual and editorial input. ESM is listed last because both first authors were supported and managed by him in his lab. Experiments were conceived and designed by MKK, PWS, AAD, and ESM. Mouse breeding, tissue preparation, IHC, data acquisition, and analysis were performed by MKK, PWS, JNH, PGG, AD, ZMW, BMF, and CJN. JDC and HRN oversaw unbiased stereological analyses. MI and JST generated pan-neuronal Bmal1-KO mice and supplied tissue from them. Intrastriatal injections were performed by AAD. MKK, PWS, AAD, and ESM wrote and edited the manuscript. All authors read and approved the final manuscript.

This work was funded by NIH training grant T32AG058518 (PWS), NIH grant K08NS101118 (AAD), and NIH grant R01AG054517 (ESM). The Washington University Genome Technology Access Center facilitated RNA-Seq experiments and is funded by NCI Cancer Center Support Grant (P30 CA91842) to the Siteman Cancer Center and by NIH ICTS/CTSA Grant (UL1TR002345). JST is an Investigator in the Howard Hughes Medical Institute.

Michael K. Kanan, Email: kananm@wustl.edu.

Patrick W. Sheehan, Email: pwsheehan@wustl.edu.

Jessica N. Haines, Email: jhaines@wustl.edu.

Pedro G. Gomez, Email: pedrogomex@wustl.edu.

Adya Dhuler, Email: adya.dhuler@wustl.edu.

Collin J. Nadarajah, Email: collin.nadarajah@wustl.edu.

Zachary M. Wargel, Email: zmw5d2@health.missouri.edu.

Brittany M. Freeberg, Email: brittany.freeberg@gmail.com.

Hemanth R. Nelvagal, Email: hemanth.nelvagal@wustl.edu.

Mariko Izumo, Email: mariko.izumo@utsouthwestern.edu.

Joseph S. Takahashi, Email: Joseph.Takahashi@UTSouthwestern.edu.

Jonathan D. Cooper, Email: cooperjd@wustl.edu.

Albert A. Davis, Email: albert.a.davis@wustl.edu.

Erik S. Musiek, Email: musieke@wustl.edu.