Drosophila were raised on standard media. Experimental crosses and F1 cohorts were kept at 26 °C under 12 h L:D cycle. Fly strains were w¹¹¹⁸ (BDSC_5905), HMS01544 (BDSC_36129),yw (BDSC_6599), yv; attP2 (BDSC 36303), elav-GAL4 (BDSC_458), ple-GAL4 (BDSC_8848), Vesicular GABA Transporter (VGAT)-GAL4 (BDSC_58980), Dopa decarboxylase (Ddc)-GAL4 (BDSC_7010), Tryptophan hydroxylase (Trh)-GAL4 (BDSC_38388), Tyrosine decarboxylase 2 (Tdc2)-GAL4 (BDSC_9313), Choline Acetyltransferase (ChAT)-GAL4 (BDSC_56500), and Vesicular glutamate transporter (VGluT)-GAL4 (BDSC_26160) from Bloomington Drosophila Stock Center. We used FlyBase to find information on the stocks listed above. The iPLA2-VIA knockout mutant ∆23 and the UAS-iPLA2-VIA-PB line were described previously (Banerjee et al., 2021).

F1 flies were collected every 2–3 days, males were grouped into cohorts of up to 12 flies, and they were transferred to new food every 5 days during aging. Climbing assays were conducted on each group at 10, 20, and 30 days from eclosion and were performed at room temperature as described in Banerjee et al. (2021). Briefly, each fly cohort was transferred to a fresh food vial and a clean empty vial was placed on top of the food vial. Flies were tapped to the bottom and given 20 s to climb 6 cm. Each assay consisted of five climbing trials. Each fly earned one point for successfully climbing to or past the 6 cm mark within 20 s. The total number of points for the group was divided by the number of flies in the group to determine the climbing index. Climbing indices were assessed for normality using Shapiro–Wilk test, and statistical comparisons were made using two-way ANOVA and Tukey’s post-hoc comparisons, performed in R. Graphs were made in R with ggplot2. Graphs show average climbing index per condition, error bars are SEM. At least 8 groups were sampled per condition. Genotypes and sample sizes are given in Table 1.

F1 flies were collected every 2–3 days without gas. After 2–3 days, F1 males of the appropriate genotype were separated over light gas and grouped into cohorts of up to 15 flies per vial. Thereafter, F1 cohorts were passed without gas to fresh vials every 2–3 days. Censors (escapers) and deaths were recorded at each pass. Adapted from Delventhal et al. (2022) and Piper and Partridge (2016). Kaplan–Meier curves were generated in R. Statistical comparison by log-rank test. Genotypes and sample sizes are given in Table 1.

iPLA2-VIA loss of function leads to reduced climbing ability after 20 days of age post-eclosion (Banerjee et al., 2021; Kinghorn et al., 2015; Iliadi et al., 2018). To determine which neurons are sensitive to the loss of iPLA2-VIA, we knocked down the gene in discrete neuronal subsets according to their neurotransmitter production. We used well-characterized GAL4 drivers that target distinct neuronal populations in the adult fly nervous system using regulatory sequences from genes involved in neurotransmitter synthesis or transport (Deng et al., 2019; Chen et al., 2013). Climbing assays were conducted at 10, 20, and 30 days post-eclosion.

We confirmed that pan-neuronal knockdown of iPLA2-VIA using the elav-GAL4 driver produced a strong climbing defect at 30 days of age, as demonstrated previously (Figure 1A, gray bars; Banerjee et al., 2021; Kinghorn et al., 2015; Iliadi et al., 2018). This effect was age-dependent, as it is for the iPLA2-VIA knockout mutant. To knock down iPLA2-VIA in only GABAergic neurons, we used the Vesicular GABA Transporter (VGAT)-GAL4 driver (Fei et al., 2010), which resulted in a strong age-dependent climbing defect that phenocopied the effect observed with pan-neuronal knockdown (Figure 1B, light blue bars).

Reports from multiple model systems have documented the sensitivity of dopaminergic neurons to iPLA2-VIA loss of function (Mori et al., 2019; Sanchez et al., 2018; Chiu et al., 2019). We therefore used the Dopa decarboxylase (Ddc)-GAL4 driver to knock down iPLA2-VIA (Li et al., 2000), which caused a small but significant age-dependent loss of climbing ability as expected (Figure 1C, light red bars). Because Ddc-GAL4 is expressed in both dopaminergic and serotonergic neurons, we established two separate sets of crosses using either pale (ple)-GAL4 to target iPLA2-VIA in dopaminergic neurons only (Friggi-Grelin et al., 2003) or the Tryptophan hydroxylase (Trh)-GAL4 driver for serotonergic neurons only (Alekseyenko et al., 2010). Knockdown in serotonergic neurons caused a similar age-dependent loss of climbing ability (Figure 1D, light purple bars), but surprisingly, knockdown in dopaminergic neurons alone did not reduce climbing ability (Supplementary Figure S1A). Thus, our data suggest that loss of iPLA2-VIA in serotonergic neurons, and likely also dopaminergic neurons, leads to age-dependent climbing defects.

In the same study, we tested an additional three neuronal subsets. We knocked down iPLA2-VIA in octopaminergic and tyraminergic neurons using Tyrosine decarboxylase 2 (Tdc2)-GAL4 (Cole et al., 2005), in cholinergic neurons using Choline Acetyltransferase (ChAT)-GAL4 (Salvaterra and Kitamoto, 2001), or in glutamatergic neurons using Vesicular Glutamate Transporter (VGluT)-GAL4 (Daniels et al., 2008). None of these manipulations caused climbing defects (Supplementary Figures S1B–D).

The above experiments showed that iPLA2-VIA is necessary in GABAergic neurons, and to a lesser extent in dopaminergic and serotonergic neurons, to maintain normal climbing ability with age. To assess whether iPLA2-VIA is sufficient in these neurons, we used GAL4 to express a UAS-driven wild-type iPLA2-VIA-PB transgene in the iPLA2-VIA^∆23 mutant, as described previously (Banerjee et al., 2021). Note that in wild-type flies with iPLA2-VIA knockdown, climbing defects were not evident until after 20 days of age. By contrast, iPLA2-VIA^∆23 mutant flies showed strong climbing defects at 20 days of age at 26 °C. This discrepancy likely is due to the pleiotropic effect of losing iPLA2-VIA from multiple vulnerable tissue types in the mutant (Banerjee et al., 2021). Nevertheless, despite the requirement for iPLA2-VIA in multiple tissue types, wild-type iPLA2-VIA expressed in the GABAergic neurons of mutant flies using the VGAT-GAL4 driver demonstrated a striking rescue of climbing ability at 20 days of age (Figure 2A, dark blue bars). Expressing wild-type iPLA2-VIA in dopaminergic neurons using the ple-GAL4 driver also improved deficient climbing of the mutant at both 15 and 20 days of age, albeit to a weaker extent than expression in GABAergic neurons (Figure 2B, dark red bars). Together, our knockdown and rescue results show that iPLA2-VIA is necessary and sufficient in GABAergic neurons for protection from age-dependent loss of climbing ability.

Loss of function iPLA2-VIA mutants have severely shortened lifespans, a common symptom of neurodegeneration (Lin et al., 2018; Kinghorn et al., 2015; Iliadi et al., 2018). Therefore, as an additional test for the importance of GABAergic neurons, we assessed whether restoring wild-type iPLA2-VIA-PB in GABAergic neurons only could rescue the lifespan of iPLA2-VIA^∆23 mutants. Notably, rescued mutant flies carrying VGAT-GAL4 and the wild-type iPLA2-VIA-PB transgene had significantly improved survival compared to mutant flies carrying just VGAT-GAL4 (Figure 2C), affirming that GABAergic neurons are a critical cell type in PLAN.

The negative geotaxis climbing assay has been a mainstay of Drosophila behavioral genetics for decades and has been used extensively to monitor neurodegeneration (Barone and Bohmann, 2013; Inagaki et al., 2010). Although numerous modifications to the assay have been developed over the years, the current study relied on a simple protocol, on account of its cost and equipment efficiencies, its utility for rapid assimilation by many experimenters, and its sufficiency in detecting age-dependent locomotor defects in iPLA2-VIA mutants (Banerjee et al., 2021). Still, subtle motor defects may have been missed in our study. It is also possible that in those conditions that failed to show an effect in our climbing assays, lower GAL4 expression resulted in lower knockdown efficacy, although the drivers we used are well-characterized and commonly used (e.g., Deng et al., 2019; Chen et al., 2013; Howard et al., 2019). Lower GAL4 expression likely explains the fact that knockdown in dopaminergic neurons only with ple-GAL4 failed to affect climbing (Supplementary Figure S1A), despite much evidence that dopaminergic neurons are sensitive to loss of iPLA2-VIA. Higher resolution techniques, preferably with computerized video tracking and image processing, should be used in the future to improve detection of altered locomotion under iPLA2-VIA loss of function conditions (Aggarwal et al., 2019; Wu et al., 2019). Furthermore, our study did not account for other types of neurological defects aside from those affecting locomotion. Of note, although survival of iPLA2-VIA mutants was improved by reintroducing the wild-type allele into GABAergic neurons only, the lifespans of the rescued flies were still markedly attenuated compared to control flies, indicating the involvement of other cell types in overall survival (Figure 2C). Single cell RNA-seq has revealed widespread iPLA2-VIA expression in the fly brain, coincident with most neurotransmitters (Davie et al., 2018), and prior studies have shown cell death throughout the fly brain in iPLA2-VIA loss of function mutants (Iliadi et al., 2018). Because our study did not examine cell death histologically, it remains to be seen precisely which brain areas and/or neuron types succumb to cell death, and over what time frame.

Drosophila and mammals share most of the major neurotransmitters, including GABA, dopamine, serotonin, acetylcholine, and glutamate, with replacement of the vertebrate epinephrine and norepinephrine with the functionally analogous octopamine and tyramine in insects (Deng et al., 2019; Deshpande et al., 2020). A wealth of studies also suggest many similar functions and behavioral outputs, including in learning and memory, circadian rhythms, reward sensing, etc. (Kasture et al., 2018). Nevertheless, the limited similarities in neuroanatomical organization between insects and vertebrates and the coexistence of multiple neuron types within each neuroanatomical region may impede direct extrapolation of findings between species. Moreover, the fact that many neurons produce multiple neurotransmitters may confound a simple classification system based on this property (Deng et al., 2019). With these caveats in mind, we turn our attention to human PLA2G6-associated neurodegeneration (PLAN).

PLAN was first described in 2006 as a collection of neurodegenerative disorders affecting children and young adults (i.e., infantile neuroaxonal dystrophy, atypical neuroaxonal dystrophy, and neurodegeneration with brain iron accumulation) with severe loss of motor coordination (Gregory et al., 2008; Morgan et al., 2006). Since then, loss of PLA2G6 in human patients has been associated also with inherited parkinsonism, as well as with dystonia and ataxia (Paisan-Ruiz et al., 2009; Guo et al., 2018; Xue et al., 2023; Erro et al., 2017). Thus, while PLA2G6 loss of function clearly causes movement disorders, it is less clear whether specific subsets of neurons are more vulnerable than others. Parkinsonism results from perturbations in the dopaminergic circuit of the basal ganglia, while perturbations in other areas, including the cerebellum, can lead to dystonia and ataxia (Dickson, 2012; Sharma, 2019). Consistently with the clinical symptoms, MRI studies of human PLAN patients have revealed abnormalities in multiple brain regions, and importantly, the most commonly affected brain region is the cerebellum, which is largely GABAergic (Gregory et al., 2008; Salih et al., 2013). Our results showing that age-dependent climbing behavior in Drosophila is strongly dependent on iPLA2-VIA in GABAergic neurons are in line with human clinical and MRI data, as well as with mouse PLA2G6 knockouts showing loss of GABAergic Purkinje cells in the cerebellum (Zhao et al., 2011). So far, Drosophila studies have utilized straightforward amorphic or hypomorphic iPLA2-VIA conditions. It has been speculated that specific molecular lesions in the gene may result in distinct clinical symptoms, but evidence has been inconclusive (Guo et al., 2018; Xue et al., 2023; Engel et al., 2010). Furthermore, it is unclear how mutations in PLA2G6 lead to neurodegeneration, although abnormalities in mitochondria, endolysosomal pathways, Ca⁺² handling, ER and presynaptic membranes, and phospholipid acyl chain composition have been noted (Mori et al., 2019; Lin et al., 2018; Kinghorn et al., 2015; Chiu et al., 2019; Zhou et al., 2016; Beck et al., 2011).

Mutations in any of the more than 20 “PARK” genes can cause parkinsonism and, in some cases, mimic sporadic PD. However, it remains unclear how closely the underlying pathological sequence of events matches between each inherited parkinsonian disorder and sporadic PD. Loss of function mutations in PLA2G6/PARK14 have been suggested to induce cellular conditions similar to those occurring in sporadic PD, e.g., by causing lysosomal dysfunction (Lin et al., 2018) or by promoting α-syn aggregation (Mori et al., 2019). Indeed, dopaminergic neurons degenerate in both PLAN and sporadic PD (Miki et al., 2017; Mori et al., 2019; Sanchez et al., 2018; Chiu et al., 2019). Serotonergic neurons also succumb in sporadic PD (Politis et al., 2012; Braak et al., 2003), and although their status in PLAN is not well-described, this study suggests they are relevant to PLAN as well. However, in sporadic PD, GABAergic regions including the globus pallidus and cerebellum largely are spared from Lewy pathology and neurodegeneration (Dickson, 2012; Braak et al., 2003). By contrast, PLAN patients often experience degeneration in the cerebellum and iron accumulation in the globus pallidus (Guo et al., 2018; Gregory et al., 2008). Our experiments indicate that GABAergic neurons are extremely sensitive to loss of iPLA2-VIA, and that this can cause severe locomotor impairment. Together, our results along with the clinical and human pathology observations may suggest that the cytopathological stimuli in PLA2G6-associated neurodegeneration are different from those in sporadic PD, although they likely eventually converge on similar or identical pathways as neurodegeneration proceeds. This is consistent with recent large scale association studies that have failed to find a link between PLA2G6 and sporadic PD (Kim et al., 2024; Tomiyama et al., 2011; Liu et al., 2020). To date, it is still unknown what determines the specific neuronal sensitivities in PD and other neurodegenerative conditions (Braak et al., 2003; Paß et al., 2021).