Ataxin2 functions via CrebA to mediate Huntingtin toxicity in circadian clock neurons

Abstract

Disrupted circadian rhythms is a prominent and early feature of neurodegenerative diseases including Huntington's disease (HD). In HD patients and animal models, striatal and hypothalamic neurons expressing molecular circadian clocks are targets of mutant Huntingtin (mHtt) pathogenicity. Yet how mHtt disrupts circadian rhythms remains unclear. In a genetic screen for modifiers of mHtt effects on circadian behavior in Drosophila, we discovered a role for the neurodegenerative disease gene Ataxin2 (Atx2). Genetic manipulations of Atx2 modify the impact of mHtt on circadian behavior as well as mHtt aggregation and demonstrate a role for Atx2 in promoting mHtt aggregation as well as mHtt-mediated neuronal dysfunction. RNAi knockdown of the Fragile X mental retardation gene, dfmr1, an Atx2 partner, also partially suppresses mHtt effects and Atx2 effects depend on dfmr1. Atx2 knockdown reduces the cAMP response binding protein A (CrebA) transcript at dawn. CrebA transcript level shows a prominent diurnal regulation in clock neurons. Loss of CrebA also partially suppresses mHtt effects on behavior and cell loss and restoration of CrebA can suppress Atx2 effects. Our results indicate a prominent role of Atx2 in mediating mHtt pathology, specifically via its regulation of CrebA, defining a novel molecular pathway in HD pathogenesis.

Fig 1. Atx2 partially suppresses mHtt induced circadian arrhythmicity and aggregation.

A. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ0 (Pdf>HttQ0 in grey) or HttQ128 (Pdf>HttQ128 in blue) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2 and attP40) and expressing two independent Atx2 TRiP RNAi lines (Atx2 TRiP #1 and #2; n = 12–41; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). Rhythmic power (P-S) is indicated for aged flies expressing HttQ128 in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP40) or expressing Atx2 TRiP RNAi (Atx2 TRiP #2, in orange). Flies were at the age of day 9–17 during DD behavior. B. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ25 or HttQ103 in PDF neurons (Pdf>HttQ25 or Pdf>HttQ103 in green) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2 and attP40) and expressing two independent Atx2 TRiP RNAi lines (Atx2 TRiP #1 and #2; n = 18–42; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). C. The number of sLNv PDF cell bodies per brain hemisphere at age day 10 is indicated for various genotypes where either Atx2 RNAi (Atx2 TRiP#2) or TRiP RNAi library control (TRiP Ctrl attP40) together without (n = 4–6) or with HttQ128 are expressed is shown (n = 19–26; *p<0.05 **p<0.01, ***:p<0.005, error bars represent standard error). D. Percentage of sLNvs (labeled with PDF in red) at age day 7 containing HttQ72-eGFP aggregates (in green) in a TRiP RNAi library control background (TRiP Ctrl attP40) and expressing a Atx2 TRiP RNAi lines (Atx2 RNAi TRiP #2) is quantified (n = 27–41; *p<0.05, **p<0.01, ***:p<0.005). E. Representative images of sLNv and lLNv for corresponding genotypes in D are shown. White arrowheads indicate lLNvs. White dot circles label sLNvs in the merged figure. Orange dash circles label sLNvs with aggregates while blue dot circles label sLNvs without aggregates in the grey scale of the green channel. Example aggregates are pointed out by orange arrows.

Fig 2. ATX2 overexpression enhances mhtt aggregation.

A. Rhythmic power (P-S) is indicated for various genotypes including flies expressing ATX2 in PDF neurons (UAS-Atx2) or in the wild-type control background (W1118) together with HttQ25 or HttQ103 (Pdf>HttQ25 or Pdf>HttQ103 n = 22–64; *:p<0.05 **p<0.01, ***:p<0.005, error bars represent standard error) B. Percentage of sLNvs (labeled with PDF in red) at age day 7 containing HttQ72-eGFP aggregates (in greee) in a wild-type background (W1118) and overexpressing ATX2 (UAS-Atx2) is quantified (n = 15–25; *p<0.05, **p<0.01, ***:p<0.005). C. Representative images of sLNv and lLNv for corresponding genotypes in B are shown. White arrowheads indicate lLNvs. White dot circles label sLNvs in the merged figure. Orange dash circles label sLNvs with aggregates while blue dot circles label sLNvs without aggregates in the grey scale of the green channel. Example aggregates are pointed out by orange arrows.

Fig 3. ATX2 lacking the PABP-binding (PAM2) domain reduces mHtt toxicity while Atx2 lacking the Lsm domain does not.

A. Rhythmic power (P-S) is indicated for various genotypes including flies expressing three independent overexpression line of ATX2 lacking PAM2 domain and one overexpression line of ATX2 lacking Lsm domain in PDF neurons (Atx2-dPAM#7/6/8 and Atx2-dLsm#9) together with HttQ0 (Pdf>HttQ0; n = 12–39) or HttQ128 (Pdf>HttQ128; n = 7–44; *:p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error) is shown. B. The number of sLNv present per brain hemisphere at day 10 is indicated for various genotypes where either ATX2 lacking PAM2 domain (Atx2-dPAM#6/7/8) or wild-type control (Ctrl) together with HttQ128 are expressed in PDF neurons is shown (n = 15–19; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). C. Representative images of sLNv and lLNv expressing HttQ46-eGFP at age day 30 are shown in the wild-type control background (Ctrl) or with ATX2-dPAM overexpression (Atx2-dPAM#6/7/8). White arrowheads indicate lLNvs without aggregates while orange arrow head indicates lLNvs with aggregates. Asterisks label sLNvs. Example aggregates are pointed out by orange arrows.

Fig 4. Atx2 reduction or ATX2 lacking the PAM domain rhythmic power improves behavioral rhythms in MJDQ78 but not mutant TDP43 models.

A. Rhythmic power (P-S) is indicated for various genotypes including flies expressing only PdfGAL4 in the wild-type control background (Pdf X Ctrl(WT)) as well as flies expressing ATX3Q78 (MJDQ78) in PDF neurons in a TRiP RNAi library control background (MJDQ78 TRiP Ctrl#2) and expressing Atx2 RNAi lines (MJDQ78 Atx2 TRiP #2) or in wild-type control background (Ctrl) and overexpression of Atx2 lacking PAM domain (MJDQ78 Atx2-dPAM#8; n = 5–17; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). B. Rhythmic power (P-S) is indicated for various genotypes including flies expressing only PdfGAL4 in the RFP expressing background (Pdf X Ctrl(RFP)) as well as flies expressing mutant TDP43 (TDP43A315T) in PDF neurons in a TRiP RNAi library control background (TDP43A315T TRiP Ctrl#2) and expressing Atx2 RNAi lines (TDP43A315T Atx2 TRiP #2) or in wild-type control (Ctrl) background and overexpression of Atx2 lacking PAM domain (TDP43A315T Atx2-dPAM#8; n = 9–15; NS:not significant, *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error).

Fig 5. Atx2 knockdown decreases Htt and mHtt levels while Atx2 or Atx2-dPAM overexpression does not affect Htt or mHtt levels prior to aggregation formation.

A. GFP Intensity in the sLNv for flies expressing HttQ25 (Q25) or HttQ46 (Q46) in a TRiP RNAi library control background (TRiP Ctrl#2) and expressing Atx2 RNAi lines (Atx2 TRiP #2) at age day 5 is quantified and shown. GFP Intensity in the sLNv without aggregates formed for flies expressing HttQ25 (Q25) or HttQ46 (Q46) in a wild-type control background (Ctrl) and expressing UAS-Atx2 and UAS-dPAM at age day 2 is quantified and shown (n = 5–38; *p<0.05, **p<0.01, ***:p<0.005). B. Representative images of LNvs (sLNv and lLNv) expressing HttQ46-eGFP at age day 2 are shown in wild-type control background (Ctrl) and overexpressing Atx2 or Atx2 lacking the PAM domain (UAS-dPAM). Blue arrowheads indicate lLNvs. Blue circles label sLNvs without aggregates. Orange dash circles label sLNvs with nuclear accumulation/aggregation of HttQ46 (which were not used for GFP intensity quantification).

Fig 6. Fmr1 knockdown partially suppresses mHtt toxicity.

A. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ0 (Pdf>HttQ0 in grey) or HttQ128 (Pdf>HttQ128 in blue) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2) and expressing two independent Fmr1 TRiP RNAi lines (Fmr1TRiP #1 and #2; n = 13–41; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). B. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ25 or HttQ103 in PDF neurons (Pdf>HttQ25 or Pdf>HttQ103 in green) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2) and expressing two independent Fmr1 TRiP RNAi lines (Fmr1TRiP #1 and #2; n = 16–42; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). C. The number of sLNv present per brain hemisphere is indicated for various genotypes where either two independent Fmr1 RNAi (Fmr1 TRiP #1/2) or TRiP RNAi library control (TRiP Ctrl) and HttQ128 are expressed is shown (n = 9–14; *p<0.05, **p<0.01, ***:p<0.005). D. Percentage of sLNvs (labeled with PDF in red) at age day 7 containing HttQ72-eGFP aggregates (in green) in a TRiP RNAi library control background (TRiP Ctrl) and expressing two independent Fmr1 TRiP RNAi lines (Fmr1 TRiP #1/2) is quantified (n = 32–65; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). E. Representative images of sLNv and lLNv for corresponding genotypes in D are shown. White arrowheads indicate lLNvs. White dot circles label sLNvs in the merged figure. Orange dash circles label sLNvs with aggregates while blue dot circles label sLNvs without aggregates in the grey scale of the green channel. Example aggregates are pointed out by orange arrows.

Fig 7. Atx2 and Fmr1 effects on HttQ128 toxicity depend on each other.

Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ0 (Pdf>HttQ0 in grey) or HttQ128 (Pdf>HttQ128 in blue) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2) and expressing two independent Fmr1 TRiP RNAi lines (Fmr1TRiP #1 and #2) or expressing both Atx2 and Fmr1 RNAi together (Atx2 TRiP#2;Fmr1 TRiP #1 and Atx2 TRiP#2;Fmr1 TRiP#2; n = 10–41; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error; n = 10–32). Some control data is reproduced from Fig 6.

Fig 8. Reduction in Atx2 but not tyf reduces peak LNv CrebA expression.

A-D. The expression level of various genes of interest (Atx2, vri, tim, CrebA) in flies expressing mGFP in the LNvs in a wild-type background (WT) or with the expression of Atx2 RNAi KK (Atx2 KD) at two time points is shown in normalized counts calculated by DEseq2. E-G. The expression level of various genes of interest (vri, tim, CrebA) in flies expressing mGFP in the LNvs in the wild-type background (WT) or in the tyf mutant (tyf(e)) at two time points is shown in normalized counts calculated by DEseq2. Asterisks indicate the significance using the adjusted p-values calculated by DEseq2 (*p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error).

Fig 9. CrebA knockdown suppresses mHtt induced arrhythmicity in two different mHtt models.

A. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ0 (Pdf>HttQ0 in grey) or HttQ128 (Pdf>HttQ128 in blue) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2) and expressing a CrebA TRiP RNAi line (CrebA TRiP #2; n = 14–34; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). Rhythmic power B. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ128 in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2) and expressing a CrebA TRiP RNAi line (CrebA TRiP#2) or a CREBA overexpression line (UAS-CrebA) or the combination of RNAi and overexpression (UAS-CrebA;CrebA TRiP#2; n = 20–34; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). C. Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ25 or HttQ103 in PDF neurons (Pdf>HttQ25 or Pdf>HttQ103 in green) in PDF neurons in a TRiP RNAi library control background (TRiP Ctrl attP2) and expressing a CrebA TRiP RNAi lines (CrebA TRiP#2; n = 19–42; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). Rhythmic power.

Fig 10. CrebA knockdown suppresses mHtt induced PDF+ cell body loss and aggregation despite elevated Htt levels.

A. The number of sLNv present per brain hemisphere at age day 10 is indicated for various genotypes where either CrebA RNAi (CrebA TRiP#2) or TRiP RNAi library control (TRiP Ctrl) without (n = 4–6) and with HttQ128 are expressed is shown (n = 20–26; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). B. Percentage of sLNvs (labelled in red) at age day 7 containing HttQ72-eGFP aggregates (in green) in a TRiP RNAi library control background (TRiP Ctrl) and expressing a CrebA TRiP RNAi lines (CrebA TRiP#2) is quantified (n = 43–44; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error). C. Representative images of sLNv and lLNv for corresponding genotypes in B are shown. White arrowheads indicate lLNvs. White dot circles label sLNvs in the merged figure. Orange dash circles label sLNvs with aggregates while blue dot circles label sLNvs without aggregates in the grey scale of the green channel. Example aggregates are pointed out by orange arrows. D. GFP Intensity in the sLNv or lLNv for flies expressing HttQ25 in a TRiP RNAi library control background (TRiP Ctrl) and expressing CrebA RNAi is quantified and shown (n = 8–20; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error).

Fig 11. CREBA overexpression partially blocks the rescue of mHtt toxicity by Atx2 knockdown.

Rhythmic power (P-S) is indicated for various genotypes including flies expressing HttQ0 (Pdf>HttQ0 in blue) or HttQ128 (Pdf>HttQ128 in red) in PDF neurons in control background (TRiP Ctrl attP2 or W1118) and expressing a Atx2 TRiP RNAi line (Atx2 TRiP#1) or a CREBA overexpression line (UAS-CrebA) or the combination of both (UAS-CrebA; Atx2 TRiP#1; n = 15–40; *p<0.05, **p<0.01, ***:p<0.005, error bars represent standard error).