As before (Melkani et al., 2013; Gill et al., 2015; Villanueva et al., 2019; Livelo et al., 2023), flies were housed together (about 10 flies per 25 mm vial) in an incubator at 25°C with constant humidity under a 12 h light:12 h dark cycle. Flies were maintained on standard Drosophila cornmeal food diet and changed to new food every 4–7 days. We used the following drivers lines: Elav(X)-Gal4 from Bloomington Drosophila Stock Center (BDSC) (BL#458) and Elav(II)-Gal4 (BL#8765). We used the following UAS lines: UAS-eGFP from BDSC (BL#5431), and HD model lines that expressed huntingtin exon 1 with 72 CAG repeats (HTTex1-PQ72-GFP) or 25 CAG repeats (HTTex1-PQ25-GFP) tagged with GFP from Dr. Norbert Perrimon (Zhang et al., 2010). Progeny of Elav drivers crossed with w¹¹¹⁸ flies were used as a driver control throughout because the lines were on a w¹¹¹⁸ background.

Fly heads were removed under a dissecting microscope and fixed with 4% PFA in PBS for 30 min at room temperature with mild agitation, then washed 3 × 10 min in 1× PBS and stored at 4°C in PBS before cryosectioning. Heads were placed in a mold with OTC and sectioned at 20 μm with a Leica CM3050 S cryostat and placed on glass slides (Fisher #15-188-48). After drying for 30 min, slides were washed 3 × 10 min with 1× PBS to remove dried OTC, then permeabilized with 1× TBST for 10 min. We blocked the slides with 3% BSA in TBST for 30 min and incubated with primary antibody for synapsin (1:500, UI Developmental Studies Hybirdoma Bank #3C11) in 1% BSA in TBST overnight at 4°C in a humidity chamber. Slides were then washed 3 × 10 with PBS and incubated for 1 h at room temperature in AlexaFluor-594 anti-mouse fluorescent secondary antibody (1:1000, ThermoFisher # A-11032). Finally, slides were washed 3 × 10 min with 1× PBS and mounted with ProLong Diamond Antifade Mountant with DAPI (ThermoFisher # P36971). After setting overnight, multichannel fluorescence images were taken with an Olympus BX63F fluorescence microscope using CellSens Software at 10× magnification to view one head section per image.

PQ-GFP aggregation was quantified using FIJI software (ImageJ v1.53v) similar to how we have quantified punctate proximity ligation signal before (Rush et al., 2020). The FITC channel images were adjusted in a batch to maximize signal-to-noise, then thresholded to exclude any background signal. ImageJ particle analyzer was run with size and circularity specified to exclude non-punctate signal on a region of interest selected around the brain, and number of puncta per section was reported.

Syanpsin immunofluorescence intensity was quantified using FIJI software (ImageJ v1.53v) similar to how we have quantified BIN1 protein levels before (Voskobiynyk et al., 2020). The TRITC channel images were adjusted in a batch to maximize signal-to-noise, then mean fluorescent intensity for each image was measured in a region of interest selected around the brain. For each experiment, we ran samples in parallel but with no primary antibody to measure background fluorescence. We measured this background fluorescence in the brain and subtracted mean background fluorescence from the fluorescence value for each image. At least two sections per fly were imaged and mean fluorescence from each fly was calculated and reported.

We conducted negative geotaxis assay as before (Ali et al., 2011; Villanueva et al., 2019). Briefly, flies were transferred into clean vials (about 10 per vial) and briefly allowed to acclimate. The vial was sharply tapped onto a pad to stimulate negative geotaxis response, and the proportion of flies the climbed to a 7 cm mark within 10 s was recorded by video for analysis. This was repeated a total of four times for each vial of flies, and the average of the four runs was calculated.

We conducted flight index assay as before (Drummond et al., 1991; Villanueva et al., 2019). Briefly, one fly at a time was released into a large plexiglass box with a light at the top. Flight index was determined by the fly’s ability to fly up (a flight index value of 6), horizontally (4), down (2), or unable to fly (0). The average flight index was calculated by dividing the sum of individual flight index values by the number of flies for each group.

Rapamycin (LC Laboratories # R-5000) was dissolved to 50 mM in ethanol, then added to food at 1:250 before pouring into vials for a final concentration of 200 μM, which is sufficient to extend fly lifespan (Bjedov et al., 2010). For vehicle control, ethanol alone was added at 1:250. To avoid developmental effects of rapamycin (Zhang et al., 2000), flies were kept on standard Drosophila cornmeal food for 1 week post-eclosion, then split into two groups and put on either rapamycin or vehicle control. Flies were kept on rapamycin or vehicle control for 3 weeks before conducting experiments at 4 weeks of age.

About 10 flies per group per cohort were flash-frozen together and stored at −80°C until use. Heads were removed and put in 150 μL Zymo Lysis Buffer and homogenized at room temperature using the Zymo Quick-RNA MicroPrep Kit (Zymo Research #R1051). RNA was purified using minicolumns and digested with DNAse I for 15 min. Next, the RNA was suspended in RNAse/DNAse free water and quantified with an Aligent Bio Tek Synergy LX Multi-mode and stored at −80°C until use. Next, we used 250 ng of RNA and 4 μL iScript RT Supermax (Bio-Rad #1708840) to synthesize cDNA. Reverse transcription was performed with 5 min of priming at 25°C, 20 min of reverse transcriptase at 46°C, then 1 min of inactivation at 95°C. We used 5 ng of cDNA for qPCR with 200 ng primers per reaction, using Ssa Advanced Universal SYBR Green Supermix (Bio-Rad #1725275) in a CFX Opus Real-Time PCR System (Bio-Rad). Expression was normalized to Rpl11 (60S ribosomal protein) and Tbp (TATA-box binding protein). PCR specificity was verified by comparing the melting curve to a predicted melting curve for the specific sequence generated by the software μMELT Quartz.¹ Results are presented as 2 − ΔΔCt values normalized to the expression of Rpl11 and Tbp and control samples. All reactions were performed in triplicate.

The following primers were used: Atg1-F: CGTCAGCCTGGTCATGGAGTA; Atg1-R: TAACGGTATCCTCGCTGAG; Hid-F: CACCGACCAAGTGCTATACG; Hid-R: GGCGGATACTGGAAGATTTGC; Pten-F: GTGCAAACGCAAACAGCCTA; Pten-R: ATCCAGTTCTGGTGGCTTCG; Ilp5-F: GCGGATTTGGATAGCTCCGA; Ilp5-R: AAAGGAACACGATTTGCGGC; Rpl11-F: CGATCTGGGCATCAAGTACGA; Rpl11-R: TTGCGCTTCCTGTGGTTCAC; Tbp-F: ATGCCCTGAGCAACATCCAC; Tbp-R: GGATCAGCGGAACCTGGTG.

Statistics and test values are described in each figure legend. 1-way or 2-way ANOVA were used, and which effect was tested is reported in each legend along with sample sizes. We used Dunnett’s post hoc to compare means to GFP mean. Data values were organized in Microsoft Excel and plotted using Graphpad Prism 9, which was also used for statistical analyses. α < 0.05 was defined as our significance threshold. Data were presented as mean ± SEM.

First, we used the UAS-Gal4 system in Drosophila (Osterwalder et al., 2001) to express different huntingtin constructs in neurons. We used the Elav-Gal4 driver on the 2nd chromosome (Elav(II)) to target expression to mature neurons (Robinow and White, 1988), and, along with a driver control crossed with wildtype W¹¹¹⁸ flies (/+), expressed three UAS constructs in neurons: GFP as an expression control, HTT exon 1 with 72 CAG repeats fused to GFP (which we refer to as Htt-PQ72), and HTT exon 1 with 25 CAG repeats fused to GFP (Htt-PQ25) (Zhang et al., 2010). We confirmed with fluorescent imaging that the GFP constructs expressed a similar amount (Supplementary Figure S1). While we did not observe any aggregation at eclosion (P0) or 1 week post-eclosion (P7), we found significant PQ-GFP aggregation in the Elav(II) > Htt-PQ72 fly brains at beginning at 2 weeks pose-eclosion (P14) and at 4 weeks post-eclosion (Figures 1A–C). We next verified whether these results were reproducible using a different Elav-Gal4 line, so we crossed the same UAS constructs with an Elav-Gal4 driver on the X chromosome (Elav(X)). In contrast to when driven by Elav(II), 45% of Elav(X) > Htt-PQ72 died before P28. To avoid survivorship bias, we measured PQ-GFP aggregation at an earlier timepoint at P21. PQ-GFP aggregated much faster in the Elav(X) > Htt-PQ72 flies than the Elav(II) > Htt-PQ72 flies, beginning by P7, and having much higher aggregation by 3 weeks (P21) than then Elav(II) > Htt-PQ72 flies had at P28 (Figures 1D–F). We interpreted that this could be due to higher expression with the Elav(X) driver than with the Elav(II) driver, which could also explain the differences in premature mortality between the drivers. Together, these results show that overexpressing mutant huntingtin in neurons causes an age- and driver-dependent aggregation of PQ-GFP in the brain.

Synaptic dysfunction is a common early feature in neurodegeneration, especially HD (Smith-Dijak et al., 2019), and synapse loss often precedes neuronal loss (Milnerwood and Raymond, 2010). Thus, we used immunohistochemistry to measure synapsin levels to determine whether Htt-PQ72 could induce changes at the synapse that are consistent with synaptic dysfunction. Synapsin is critical for neurotransmitter vesicle release (Cesca et al., 2010) and binds mutant huntingtin in the presynaptic compartment to inhibit its phosphorylation (Xu et al., 2013), which predicts dysfunction in HD models (Liévens et al., 2002). We found that at 3 weeks of age, Elav(X) > Htt-PQ72 flies had reduced synapsin levels in the brain relative to GFP, while Elav(X) > Htt-PQ25 flies did not (Figures 2A,C). However, we did not see a significant change in synapsin staining in any flies using the Elav(II) driver at 4 weeks of age (Figures 2B,D). To determine whether synapsin loss occurs with PQ-GFP aggregation or after significant aggregation, we stained for synapsin in P7 Elav(X) flies, when PQ-GFP aggregation has already occurred (Figure 1A), and found that there was no difference in synapsin staining (Figures 2E,F). Thus, PQ-GFP aggregation precedes synapsin loss in the brain of Elav(X) > Htt-PQ72 flies, which occurs with age.

HD is characterized by muscle dysfunction, and we have previously found that Htt-PQ72 causes dysfunction in cardiac and skeletal muscle (Melkani et al., 2013; Barwell et al., 2023). It has been proposed that this peripheral muscle dysfunction is primarily caused by peripheral expression of mutant huntingtin, and that peripheral mutant huntingtin could cause neuronal dysfunction in the brain (Chuang and Demontis, 2021). We wanted to complement this idea and test the converse – whether neuronal huntingtin can disrupt locomotor performance independent from expression in the muscle. To do so, we conducted two assays to measure the performance of two types of skeletal muscle: negative geotaxis assay, which measures climbing muscle performance, and flight index assay, which measures flight muscle performance (Villanueva et al., 2019). We chose to use the Elav(II) driver to allow for observing more subtle phenotypes induced by lower aggregation or expression of constructs. While there were no significant changes in climbing muscle performance at P7, before PQ-GFP aggregation, or at P14, when aggregation is beginning (Figure 1A), Elav(II) > Htt-PQ72 flies displayed progressive dysfunction in climbing muscle performance starting at P21, beyond the loss of muscle performance associated with normal aging (Figure 3A). There was also some climbing muscle dysfunction in Elav(II) > Htt-PQ25 flies at P28, which is not surprising since some dysfunction has been reported by overexpression of shorter PQ chains with aging in vivo (Morley et al., 2002). To determine if neuronal mutant huntingtin causes general muscle performance loss, we performed a flight index assay to measure flight muscle performance. To our surprise, we found no change in flight muscle performance at P28 (Figure 3B). As it is possible that higher PQ-GFP in neurons could decrease flight muscle performance, we repeated the flight index assay using the Elav(X) driver. Again, we observed no change in flight muscle performance at 3 weeks of age (Figure 3C). Together, these results show that neuronal Htt-PQ72 can decrease climbing muscle performance in an age-dependent manner without decreasing flight muscle performance.

Autophagy plays an important role in aging and metabolism, and has been implicated in many neurodegenerative diseases including HD (Croce and Yamamoto, 2019). Rapamycin stimulates autophagy and prevents some mutant huntingtin-induced dysfunction (Djajadikerta et al., 2020), so we next tested whether it impacted PQ-GFP aggregation in the brain. We placed 1 week-old flies on food supplemented with 200 μM rapamycin or vehicle control food for 3 weeks. First, we verified the mechanism of action of rapamycin by isolating heads and conducting PCR for genes associated with autophagy and TOR signaling (Figure 4A). We found that rapamycin significantly increases expression of Atg1, which is responsible for the formation of autophagosomes (Mizushima, 2010; Stjepanovic et al., 2014). Additionally, we measured expression of Hid, which can induce autophagy (Baehrecke, 2003; Juhász and Sass, 2005), and pten, which is negatively associated with TOR activity (Manning and Cantley, 2007; Das et al., 2012), and found them increased by rapamycin, demonstrating that rapamycin treatment led to inhibition of TOR as expected. As an independent measure of TOR activity, we measured ilp5 expression, which is positively regulated by TOR and is important for nutrient signaling and metabolism (Texada et al., 2022; Ling and Raikhel, 2023). Rapamycin significantly decreased ilp5, again demonstrating that rapamycin inhibited TOR signaling (Figure 4B). Next, we determined whether rapamycin prevented PQ-GFP aggregation and other mHTT-associated dysfunction. We found that rapamycin reduced PQ-GFP aggregation, though it did not prevent a loss of synapsin in aged Elav(X) flies (Figures 4C–E). Next, we determined whether rapamycin was sufficient to prevent PQ-GFP aggregation and locomotor performance loss in Elav(II) flies. We found that rapamycin again reduced PQ-GFP aggregation and also prevented a loss of locomotor performance in Elav(II) > Htt-PQ72 flies (Figures 4F–H). There was a decrease in climbing performance in the control flies relative to prior experiments (Figure 3), likely due to chronic ethanol exposure (Chvilicek et al., 2020; Nuñez et al., 2023) as ethanol was used to dissolve the rapamycin and was included in the control food as a vehicle. Altogether, these results show that rapamycin is sufficient to prevent aggregation of mutant huntingtin in the brains of flies and can prevent the loss of locomotor performance.