Expression of mutant CHMP2B linked to neurodegeneration in humans disrupts circadian rhythms in Drosophila

Abstract

Mutations in CHMP2B, an ESCRT-III (endosomal sorting complexes required for transport) component, are associated with frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). Neurodegenerative disorders including FTD are also associated with a disruption in circadian rhythms, but the mechanism underlying this defect is not well understood. Here, we ectopically expressed the human CHMP2B variant associated with FTD (CHMP2B^Intron5) in flies using the GMR-GAL4 driver (GMR>CHMP2B^Intron5) and analyzed their circadian rhythms at behavioral, cellular, and biochemical level. In GMR>CHMP2B^Intron5 flies, we observed disrupted eclosion rhythms, shortened free-running circadian locomotor period, and reduced levels of timeless (tim) mRNA-a circadian pacemaker gene. We also observed that the GMR-GAL4 driver, primarily known for its expression in the retina, drives expression in a subset of tim expressing neurons in the optic lobe of the brain. The patterning of these GMR- and tim-positive neurons in the optic lobe, which appears distinct from the putative clusters of circadian pacemaker neurons in the fly brain, was disrupted in GMR>CHMP2B^Intron5 flies. These results demonstrate that CHMP2B^Intron5 can disrupt the normal function of the circadian clock in Drosophila.

Keywords: CHMP2B; ESCRT; circadian Rhythms; endosomal Lysosomal Pathway.

Figure 1

Composite double‐plotted actograms of 1‐3‐day‐old (A) wild type (w¹¹¹⁸), (B) GMR‐GAL4, (C) GMR>CHMP2B^Intron5 flies, and (D) a scatter plot of the free‐running locomotor period of the three genotypes show that GMR‐driven expression of CHMP2B^Intron5 causes a significant reduction in the free‐running locomotor rhythm (wild type—23.82 ± 0.03 hours (n = 29); GMR‐GAL4—23.67 ± 0.06 hours (n = 25); GMR>CHMP2B^Intron5—23.11 ± 0.07 hours (n = 37); error bars—standard error of measurement; n.s.—not significant; ***—P < 0.001; ANOVA, Bonferroni pairwise comparison). The timing of 12:12 LD and DD cycles is shown by alternating black and white rectangles on top of the actograms and by light gray shading on the actograms, respectively

Figure 2

Histograms plotting the eclosion percentage in 1 hour bins of (A) wild type (Canton‐S), (B) per⁰, (C) GMR‐GAL4, and (D) GMR>CHMP2B^Intron5 flies maintained in 12 hour light‐12 hour dark (LD) cycle. The wild type and GMR‐GAL4 flies preferentially eclose around light on‐off transition (n_wt = 100, n_{GMR ‐Gal4} = 59). The per⁰ flies lack robust circadian preference for eclosion (P < 0.01; chi‐squared test; n_per0 = 67). The GMR>CHMP2B^Intron5 flies show a significant dampening of the eclosion peak around light on‐off transition (P < 0.01; chi‐squared test; n = 77). The timing of 12 hour light‐12 hour dark (LD) cycle is shown by alternating black and white rectangles on top of the histograms

Figure 3

Line charts plotting the (A) development rate and (B) longevity show that the GMR>CHMP2B^Intron5 flies have similar pupation and eclosion rates compared to that of wild type (Canton‐S) flies (wild type—n = 7 batches of 100 larvae each and GMR>CHMP2B^Intron5—n = 5 batches of 100 larvae each; error bars—standard error of measurement; pupation—P = 0.07; eclosion—P = 0.49; Welch's t test). The GMR>CHMP2B^Intron5 flies have a similar lifespan compared to that of wild type flies (wild type—n = 100 and GMR>CHMP2B^Intron5—n = 46; P = 0.8; Wilcoxon rank test)

Figure 4

Representative images of whole mount brains of (A) GMR>GFP, (B) GMR>GFP + CHMP2B^Intron5, and (C) GMR>GFP + Hid show that GMR‐driven expression of CHMP2B^Intron5 causes disruption of patterning in the band of GMR‐positive cells in optic lobe. The band of GMR‐positive cells in the optic lobe and the size of optic lobe is more severely damaged in the GMR>GFP + Hid flies compared to that of GMR>GFP + CHMP2B^Intron5 flies. Large dotted lines outline the brain while small dotted lines demarcate the optic lobe/central brain boundary. Arrows in (A) and (B) mark the band of large cells in optic lobe; scale bar: 50 μm

Figure 5

Representative images of whole mount brains of (A) GMR>GFP and (B) tim‐DsRed show that GMR‐driven expression of GFP in a band of cells in the optic lobe partially overlaps with tim‐driven expression of DsRed as shown in the (C) merged image. Arrows highlight the relevant band of cells with substantial overlap of DsRed and GFP signal, inset is a digital zoom of the area marked by arrows; scale bar: 50 μm

Figure 6

Scatter plots of the normalized transcript levels of (A) tim and (B) per at ZT6 and ZT12 show that the GMR>CHMP2B^Intron5 flies have a significant reduction in tim transcript levels at ZT12 when compared to wild type (Canton‐S) flies (*—P < 0.05; Welch's t test; error bars: standard error of measurement; n.s.—not significant; a.u.—normalized arbitrary unit; n = 5‐7—wild type and GMR>CHMP2B^Intron5). Both wild type and GMR>CHMP2B^Intron5 flies show significant increase in tim transcript levels (ZT12:ZT6 fold change_tim—wild type = 2.9; GMR>CHMP2B^Intron5 = 2.6 (two‐way ANOVA for time—P = 0.0003 and genotype—P = 0.003). However, the per transcript levels show a trend for increase between ZT6 and ZT12 in wild type and GMR>CHMP2B^Intron5 flies (ZT12:ZT6 fold change_per—wild type = 2.3; GMR>CHMP2B^Intron5 = 1.6; two‐way ANOVA for time—P = 0.11 and genotype—P = 1.0)