Circadian clock disruption promotes the degeneration of dopaminergic neurons in male Drosophila

Abstract

Sleep and circadian rhythm disruptions are frequent comorbidities of Parkinson's disease (PD), a disorder characterized by the progressive loss of dopaminergic (DA) neurons in the substantia nigra. However, the causal role of circadian clocks in the degenerative process remains uncertain. We demonstrated here that circadian clocks regulate the rhythmicity and magnitude of the vulnerability of DA neurons to oxidative stress in male Drosophila. Circadian pacemaker neurons are presynaptic to a subset of DA neurons and rhythmically modulate their susceptibility to degeneration. The arrhythmic period (per) gene null mutation exacerbates the age-dependent loss of DA neurons and, in combination with brief oxidative stress, causes premature animal death. These findings suggest that circadian clock disruption promotes dopaminergic neurodegeneration.

Fig. 1. Rhythmic vulnerability of PAM neurons to oxidative stress.

a A schematic of the cell bodies of DA neuron clusters in the adult fly brain. PAM protocerebral anterior medial, PAL protocerebral anterior lateral, PPL1 and 2 protocerebral posterior lateral 1 and 2, PPM1, 2, and 3 protocerebral posterior medial 1, 2, and 3. b DA neuron counts in w¹¹¹⁸ flies per hemisphere in each cluster 7 days after a 4-h 10% H₂O₂ treatment performed at ZT20. The control group was treated with water only. DA neurons were detected by anti-TH immunostaining. Neurodegeneration was observed only in the PAM cluster following the H₂O₂ treatment (n = 18 hemispheres analyzed for both groups). ***p < 0.0001 (two-tailed t-test, comparing the control and H₂O₂-treated groups). In box plots in this and all following figures, box boundaries are the 25th and 75th percentiles, the horizontal line across the box is the median, and the whiskers indicate the minimum and maximum values. The dots represent all data points. c A schematic of the circadian H₂O₂ treatment protocol. d, e Quantification of the number of PAM neurons after the 4-h 10% H₂O₂ treatment performed at different times in LD (control, n = 20; ZT0, 4, 8 and 16 n = 15 each; ZT12, n = 14; ZT16, n = 17 hemispheres) (d) or DD (control, n = 16; CT0, n = 18, CT4, n = 17; CT8 and 20, n = 15 each; CT12 and 16, n = 16 each) (e). The x-axis indicates the times when H₂O₂ was applied. The control group was treated with water only at ZT20 in LD (d) and CT20 in DD (e). At all time points, PAM neuron counts in the H₂O₂ treatment group are significantly smaller than those in the control group. ****p < 0.0001 (one-way ANOVA with Tukey’s post hoc test). Within the H₂O₂-treated group, flies treated at ZT20 in LD (d) and CT4 in DD (e) showed significantly greater cell loss than the treatment at any other time. Different lowercase letters represent statistical significance by one-way ANOVA with Tukey’s HSD post hoc test. Source data are provided as a Source Data file.

Fig. 2. The clock gene per controls the magnitude and rhythms of the vulnerability of PAM neurons to oxidative stress.

a Representative maximum-projection images of PAM neurons in w¹¹¹⁸ and per⁰ flies at indicated ages. PAM neurons were visualized by anti-TH antibodies (green) and RedStinger driven by the R58E02 driver (magenta). Approximately half of the neurons in the PAM cluster are visible. Scale bar, 20 µm. b, c The number of PAM neurons detected by anti-TH immunostaining (w¹¹¹⁸ day1, n = 33; day7, n = 21; day14, n = 18 hemispheres. per⁰ day1, n = 31; day7, n = 52; day14, n = 21.) (b) and by the expression of R58E02-driven RedStinger (w¹¹¹⁸ day1, n = 12; day7, n = 17; day14, n = 27 hemispheres. per⁰ day1, n = 13; day7, n = 23; day14, n = 26.) (c). per⁰ flies display developmental and age-dependent loss of PAM neurons. **p < 0.01 and ****p < 0.0001 (two-tailed t-test or two-tailed Mann–Whitney U test). d PAM neuron counts in Canton-S, w¹¹¹⁸, and per⁰ were assessed by anti-TH staining 7 days after the treatment with different concentrations of H₂O_2. Treatment was performed from ZT1 for 4 h. (Canton-S control, n = 16, 2.5% H₂O₂, n = 12; 5%, n = 15; 10%, n = 16 hemispheres. w¹¹¹⁸ control, n = 17; 2.5%, n = 16; 5%, n = 16; 10%, n = 17. per⁰ control, n = 18; 2.5%, n = 18; 5%, n = 17; 10%, n = 11.) per⁰ flies display increased PAM neuron susceptibility compared to control genotypes. *p < 0.05, ***p < 0.001, and ****p < 0.0001 (one-way ANOVA with Dunnett’s multiple comparisons test). e, f PAM neuron counts in per⁰ flies after a 4-h 5% H₂O₂ treatment were performed at different time points in LD (control, n = 31; ZT0, n = 18; ZT4, n = 12; ZT8, n = 13; ZT12, n = 16; ZT16, n = 12; ZT20, n = 16 hemispheres) (e) or DD (control, n = 30; CT0, n = 15; CT4, n = 13; CT8, n = 14; CT12, n = 15; CT16, n = 14; CT20, n = 15 hemispheres) (f). The x-axis indicates the time points when H₂O₂ was applied. The control group was treated with water only at ZT0 in LD (e) and CT0 in DD (f). At all time points, PAM neuron counts in the H₂O₂ treatment group were significantly smaller than those in the control group. *p < 0.05 (one-way ANOVA with Tukey’s HSD post hoc test). Within the H₂O₂-treated group, flies treated at ZT12 in LD (e) displayed a significantly greater cell loss than at any other time point. No difference was observed between time points in DD (f). Different lowercase letters represent statistical significance by one-way ANOVA with Tukey’s post hoc test. Source data are provided as a Source Data file.

Fig. 3. Oxidative stress elicited by H₂O₂ ingestion but neither feeding rhythms nor dehydration causes circadian vulnerability of PAM neurons.

a Feeding patterns of w¹¹¹⁸ flies in LD (left) and DD (right). Flies were fed for 4 h with 10% H₂O₂ solution containing blue dye. The control solution contained only water and blue food dye. n = 15–25 flies. The x-axis indicates the time points when the flies started to be fed. The y-axis represents the relative absorbance per ten flies. The dots represent the values of three independent experiments. Feeding levels in the control group at ZT8 were significantly higher than at other time points in LD. *p < 0.05 in any pairwise comparison (ANOVA with Tukey’s post hoc test). Error bars indicate the range. No significant difference was observed between time points in all other groups. b PAM neuron counts after 9 h of food and water deprivation, started at ZT15. The control group was deprived of food and water for 5 h, followed by 4 h of water access. n = 15 hemispheres for both groups. No significant difference between groups (t-test). c, d Catalase was expressed in PAM neurons with the R58E02 driver, and its effect on H₂O₂-induced PAM neuron loss was assessed by anti-TH immunostaining in LD (c) and DD (d). Catalase expression prevented PAM neuron loss following a 4-h 10% H₂O₂ treatment performed at ZT20 (c), as well as at CT4 and CT16 (d). The control group was treated with water at the same time point. ***p < 0.001 (p = 0.0004 in (c); p = 0.0004 between R58E02 control and CT16 H₂O₂, p = 0.0006 UAS-Catalase control vs. CT16 H₂O₂ in (d)) and ****p < 0.0001 (Kruskal–Wallis test with Dunn’s multiple comparisons test). In (c), R58E02 control, n = 22; H₂O₂, n = 54 hemispheres. UAS-Catalase control, n = 41; H₂O₂, n = 55. R58E02>Catalase, n = 50; H₂O₂, n = 37. In (d), R58E02 control, n = 20; CT4 H₂O₂, n = 25; CT16 H₂O₂ n = 52 hemispheres. UAS-Catalase control, n = 43; CT4 H₂O₂, n = 68; CT16 H₂O₂, n = 49. R58E02>Catalase control, n = 39; CT4 H₂O₂, n = 47; CT16 H₂O₂, n = 27. e ROS levels within the PAM neurons were measured using MitoSOX red in flies expressing EGFP with the R58E02 driver in LD (left) and DD (right). ZT2, n = 11; ZT 8, 14 and 20; 8 flies. CT2, 14 and 20, n = 9; CT8, n = 7. ROS levels are significantly elevated at ZT20 in LD but do not differ among time points in DD (ANOVA with Tukey’s post hoc test). Source data are provided as a Source Data file.

Fig. 4. PER expression in clock neurons preserves PAM neurons.

a–c Genetic rescue of per in clock neurons in per⁰ mutants. PAM neurons were counted using anti-TH staining 7 days after a control treatment with water (a) and after a 4-h treatment with 2.5% H₂O₂ performed at ZT12 in LD (b). In (c), the control group was treated with water at CT16, and 2.5% H₂O₂ treatment was performed at CT4 and CT16. Genotypes are indicated on the x-axis. a > b represents that UAS-transgene b is driven by the GAL4 driver a. n = 11–24 hemispheres (see Source Data for individual sample numbers). Rescue genotypes significantly improve the preservation of PAM neurons after the H₂O₂ treatment. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 (one-way ANOVA with Dunnett’s multiple comparisons test, or Kruskal–Wallis test with Dunn’s multiple comparisons test). All the rescue genotypes in (c) exhibit a significantly higher number of PAM neurons compared to per⁰ (****p < 0.0001 for each pairwise comparison, not shown in the figure for clarity). n = 11–19 (a), n = 11–24 (b), and n = 20–39 hemispheres (c). d trans-Tango experiments using DvPdf-GAL4 and Pdf-GAL4 combined with TH immunostaining (green). Several postsynaptic targets of DvPdf neurons (magenta) were identified as PAM neurons, whereas no postsynaptic signal of Pdf neurons was found within the PAM cluster. Presynaptic signals by DvPdf-GAL4 and Pdf-GAL are not visible in these pictures. Two independent experiments. Scale bar, 20 µm. e A schematic representation of the trans-Tango labeling results. Projections of the l-LNvs are not shown for clarity. DvPdf-GAL4- positive but Pdf-GAL4-negative neurons, i.e., LNds, are presynaptic to a subset of PAM neurons. Source data are provided as a Source Data file.

Fig. 5. Identification of the vulnerable subpopulations of PAM neurons.

a, b Quantification of the number of neurons labeled by GFP driven with different PAM neuron drivers, 7 days after a 4-h 10% H₂O₂ treatment at ZT20. The controls were treated with water only. a UAS-EGFP was driven with classical GAL4 drivers as indicated in the x-axis. Among the drivers tested, only the R58E02-labeled neurons include a subpopulation vulnerable to oxidative insults. **p = 0.035 (two-tailed t-test). n = 7–25 hemispheres (see Source Data for individual sample numbers). b PAM subpopulations were labeled with split-GAL4 drivers. Neurons labeled by MB299B, MB441B, and MB315C contain vulnerable PAM subpopulations. ***p < 0.001 (p = 0.0009 MB441B control vs. H₂O₂; p = 0.0004 MB315C control vs. H₂O₂) and ****p < 0.0001 (two-tailed t-test). n = 8–98 hemispheres (see Source Data for individual sample numbers). c Summary of the results from (a) and (b). The mean number of cells labeled by the given driver and of cells lost following the H₂O₂ treatment are shown. The numbers in the brackets indicate the range of values. AD activation domain, DBD DNA-binding domain. d Schematic representation of PAM subpopulations vulnerable to oxidative stress. PAM-α1, -γ5, and -γ3 neurons are labeled by MB299B, MB315C, and MB441B, respectively, with three PAM-α1 neurons co-expressing MB299B and MB315C (left). H₂O₂ treatment selectively degenerates approximately half of the PAM-α1 neurons and a few cells each from the PAM-γ5 and -γ3 subgroups (right). e, f 4-h 10% H₂O₂ treatment was performed at ZT8 and ZT20, and the number of MB299B-positive neurons was quantified 7 days post-treatment. e Representative confocal images of MB299B>EGFP neurons. Scale bar, 20 µm. f Both experimental groups have significantly fewer remaining neurons than the control group treated with water only at ZT20. H₂O₂ treatment at ZT20 caused a greater loss of neurons than at ZT8. *** p < 0.001 (p = 0.0001 control vs. ZT8 H₂O₂; p = 0.0003 ZT8 vs. ZT20 H₂O₂) and **** p < 0.0001 (two-tailed t-test). Control, n = 37; ZT20 H₂O₂, n = 98, ZT8 H₂O₂, n = 14 hemispheres. g In DD, 10% H₂O₂ treatment performed at CT4 and CT16 both led to a significant loss of MB299B neurons compared to the water-only control treatment at CT4. ****p < 0.0001 (two-tailed Mann–Whitney test). The number of remaining MB299B neurons was similar between CT4 and CT16 treatments. Control, n = 13; CT4 H₂O₂, n = 55, CT16 H₂O₂, n = 71 hemispheres. Source data are provided as a Source Data file.

Fig. 6. MB299B neurons display Ca²⁺rhythms.

a Projections of the LNds contact dendritic arbors of MB299B neurons. Image created using the hemibrain connectome data via the NeuPrint tool. b Illustration of the method for live GCaMP imaging. c, d MB299B > GCaMP7s fluorescence levels in MB299B neurons throughout 24 h on the first day in DD following LD-entrainment in w¹¹¹⁸ (c) and per⁰ (d). Relative fluorescence intensity (mean ± SEM) was determined from three independent experiments. In (c), the analysis included an average of 79 (53–146 cells) cells per timepoint from 5–10 w¹¹¹⁸ flies (see Source Data for individual sample numbers per timepoint). In (d), an average of 76 (44–131) cells per timepoint from 5–10 per⁰ flies were analyzed (see Source Data for individual sample numbers per timepoint). Source data are provided as a Source Data file.

Fig. 7. A single short-term H₂O₂ treatment increases nighttime sleep.

Seven-day-old w¹¹¹⁸ flies were treated with 10% H₂O₂ or water for 4 h starting at ZT20 in LD and then placed in the activity monitor. The sleep and activity of flies from age day 11 to 13 (d11-13) (control, n = 38; H₂O₂, n = 40 flies) and day 17 to 19 (d17-19) (control, n = 33; H₂O₂, n = 29 flies) were analyzed. a, b Total sleep duration over 24 h (24h), during daytime (light period, LP) and the night (dark period, DP) in water-only control and H₂O₂-treated flies from day 11 to 13 (a) and day 17 to 19 (b). H₂O₂-treated flies show an increase in total sleep in both age groups. *p < 0.05 and **p < 0.01 (two-way ANOVA with Šídák’s multiple comparisons test). c, d Mean sleep episode duration from day 11 to 13 (c) and day 17 to 19 (d). H₂O₂-treated flies show an increase in nighttime sleep episode duration in both age groups. **p < 0.01 and ****p < 0.0001 (two-way ANOVA with Šídák’s multiple comparisons test). e, f Mean activity counts during the wake period from day 11 to 13 (e) and day 17 to 19 (f). No significant difference was observed between the control and treated groups (two-way ANOVA with Šídák’s multiple comparisons test). Source data are provided as a Source Data file.

Fig. 8. A single short-term H₂O₂ treatment causes premature death in per⁰ flies.

PAM neuron counts and survival of w¹¹¹⁸ and per⁰ flies following a 4-h exposure to 10% H₂O₂ at ZT20 when flies were 7 days old. a, b PAM neuron counts across aging in w¹¹¹⁸ (a) and per⁰ (b). PAM neuron loss was observed as early as day 14 in both genotypes compared to control groups treated with water only. Neurodegeneration progresses with age but is not accelerated in H₂O₂-treated flies. *p = 0.02 and ***p < 0.001 (two-tailed t-test). n = 12–20 hemispheres (see Source Data for individual sample numbers). Since H₂O₂-per⁰ flies exhibited a high mortality rate after 28 days, PAM neurons of this group were not examined thereafter. c, d % of surviving flies in w¹¹¹⁸ (c) and per⁰ (d). H₂O₂ treatment did not affect the survival of w¹¹¹⁸ flies, whereas H₂O₂-treated per⁰ flies had significantly reduced lifespans compared to the water-treated control group. ***p = 0.0002 (Log-rank test). In (c), n = 3 independent experiments for both control and H₂O₂ treatment, with a total of n = 324 w¹¹¹⁸ flies for control and n = 262 for H₂O₂ treatment. In (d), n = 3 independent experiments for both the control and the H₂O₂-treated groups, with a total of n = 190 per⁰ flies for the control and n = 203 for the H₂O₂-treated group. Source data are provided as a Source Data file.