INAD patient-derived skin fibroblasts have been used in several studies to explore the molecular mechanisms of INAD (Davids et al., 2016; Kinghorn et al., 2015; Sun et al., 2021; Villalón-García et al., 2022). We tested the specificity of three commercially available PLA2G6 antibodies and found one antibody that recognizes PAL2G6 (sc-376563 Santa Cruz Biotechnology; Figure 1—figure supplement 1A-B). This antibody recognizes a band of the proper molecular weight in HEK-293T cells as well as induced pluripotent stem cells (iPSCs) and neural progenitor cells (NPCs) derived from a healthy person (Figure 1—figure supplement 1A and Figure 1A). Moreover, this band was absent in HEK-293T cells that express PLA2G6 shRNAs and iPSCs and NPCs derived from an INAD patient with a PLA2G6-R70X variant (29-3; see below) (Figure 1—figure supplement 1B and Figure 1A). These data show that this antibody can specifically recognize endogenous PLA2G6.

To explore the expression levels of PLA2G6 in skin fibroblasts, we used skin fibroblasts derived from a healthy person as control. We observed very low expression levels of PLA2G6 (Figure 1—figure supplement 1C). We then obtained skin fibroblasts from two INAD patients and their parents, labeled Families 1 and 2 (Figure 1—figure supplement 1—source data 1), and observed that all six skin fibroblast lines, including four from unaffected parents and two from INAD patients, express no or very low levels of PLA2G6 (Figure 1—figure supplement 1D).

We previously showed that knocking down PLA2G6 in Neuro-2A cells leads to the expansion of lysosomes. Moreover, loss of iPLA2-VIA in flies leads to the disruption of mitochondrial morphology (Lin et al., 2018). To assess the phenotypes associated with the patient-derived skin fibroblasts, we determined the levels of LAMP2, a lysosomal marker, as well as the morphology of mitochondria. The patient-derived skin fibroblasts (88101) from Family 1 show an elevation of LAMP2 levels and a disrupted mitochondrial morphology (Figure 1—figure supplement 1D–E). However, the patient-derived skin fibroblasts (1914560) from Family 2 do not show any of the phenotypes mentioned above (Figure 1—figure supplement 1D–E). Hence, the phenotypes are inconsistent in patient-derived skin fibroblasts. In summary, skin fibroblasts express no or very low levels of PLA2G6 and exhibit highly variable phenotypes making it difficult to interpret data derived from skin fibroblasts. We, therefore, opted to use iPSCs and neurons derived from these cells to characterize phenotypes associated with INAD.

To assess the phenotypes associated with INAD patient-derived cells, we reprogramed lymphoblasts from a patient in Family 2 (PLA2G6-R70X) into iPSCs. We obtained two clones, 29-1 and 29-3. As mentioned above, 29-3 was used in Figure 1A to determine the expression of PLA2G6. To generate a control for the INAD patient-derived iPSCs, we used CRISPR technology to correct the variant in patient-derived iPSCs (29-1) to generate 29-2 (Figure 1—figure supplements 2 and 3). Hence, 29-1 and 29-2 are an isogenic pair of INAD patient-derived iPSCs. Patient and isogenic gene-corrected control lines were differentiated into ventral midbrain dopaminergic (DA) neurons. Each line uniformly expressed the positional transcription factor signature markers for ventral midbrain floor plate region neural progenitor cells (NPC) (OTX2+, LMX1A+, FOXA2+, NESTIN+, and are negative for the forebrain marker FOXG1; Figure 1—figure supplement 2B–C; Nolbrant et al., 2017). Upon further differentiation, these NPCs gave rise to a robust population of ventral midbrain DA neurons (TH⁺) with typical neuronal morphologies (Figure 1—figure supplement 2D). We then used the isogenic pair (29-1 and 29-2) of INAD patient-derived ventral midbrain NPCs as well as DA neurons differentiated from these NPCs in subsequent experiments.

Given that variants in PLA2G6 also cause parkinsonism, we explored if there are defects associated with INAD patient-derived DA neurons. As shown in Figure 1—figure supplement 3B, we did not observe obvious overall cell morphological changes in INAD patient-derived (29-1 and 29-3) and genetically corrected DA neurons (29-2). We measured the levels of Glucosylceramide (GlcCer), and examined the morphology of lysosomes (LAMP2) and mitochondria (ATP5a). GlcCer is easily observed in both control- and patient-derived DA neurons (arrows in Figure 1—figure supplement 3C; a-c) but not in the undifferentiated NPCs (arrowheads in Figure 1—figure supplement 3C; a-c). This suggests that DA neurons, but not NPCs generate high levels of GlcCer. However, the GlcCer levels are two to three fold higher in INAD patient-derived DA neurons (29-1 and 29-3) than in corrected cells (29-2) (Figure 1B and Figure 1—figure supplement 3C; a-c). Furthermore, lysosomes are increased in number and enlarged in size in cells that carry the variant compared to corrected cells (Figure 1C and arrows in Figure 1—figure supplement 3C; d-f). Finally, we also observed enlarged mitochondria in the INAD patient-derived DA neurons (29-1 and 29-3) when compared to corrected cells (29-2) (Figure 1D and Figure 1—figure supplement 3C; g-l).

To extend our observations, we also examined the changes of lysosomal and mitochondrial morphology in NPCs. As shown in Figure 1E, we observed an increase in size and number of lysosomes in the INAD patient-derived NPCs (Figure 1E; arrows in upper panel; quantifications on the right). Moreover, the mitochondria in the corrected NPCs form a connected network (Figure 1E; arrowheads in lower panel). However, the mitochondria in the INAD patient-derived NPCs are enlarged and fragmented (Figure 1E; arrows in lower panel; quantifications on the right). To assess the levels of PLA2G6 and LAMP2, we performed western blots on NPCs (Figure 1F; quantified in the right panels). Both clones of the INAD patient-derived NPCs, 29-1 and 29-3, do not express PLA2G6 (Figure 1F; lanes 1 and 3, respectively). However, the genetically corrected cells obviously express PLA2G6 (Figure 1F; lane 2). Interestingly, LAMP2 levels are significantly upregulated in INAD patient-derived NPCs compared to corrected cells (Figure 1F), consistent with a lysosomal expansion. To assess the function of the mitochondria, we measured ATP levels in 29-1, 29-2, and 29-3 patient-derived NPCs. We found that the ATP levels are significantly higher in the genetically corrected cells (29-2) than in the other two lines, indicating defective mitochondrial function in the patient-derived NPCs (Figure 1G).

Next, we performed lipidomic assays to measure the levels of Ceramides and their derivatives in the INAD patient-derived NPCs (29-1) and the genetically corrected cells (29-2). As shown in Figure 1H, we found a significant increase in total Cer and HexCer levels and they are about two fold higher in the patient-derived NPCs (Figure 1H; left panel and Figure 1—source data 3). However, although the levels of Sphinganine and Sphingosine are increased in the patient-derived NPCs, they are not statistically significant (Figure 1H; right panel and Figure 1—source data 3). The data show that INAD patient-derived NPCs exhibit an elevation of the HexCer (including GlcCer) and many different other types of ceramides. In summary, loss of PLA2G6 leads to an elevation of GlcCer, expansion of lysosomes, and defects in mitochondrial dynamics and functions. Moreover, these phenotypes are reversed in genetically corrected cells, indicating a causative relationship.

To study the neuropathology in mice, we obtained two INAD mouse models: mice that lack Pla2g6 (Pla2g6^KO/KO) (Bao et al., 2004) and mice that carry a homozygous G373R point mutation (Pla2g6^G373R/G373R) (Wada et al., 2009). These two models exhibit similar neuropathological defects, as explained in the introduction, but the severity of the phenotypes is much stronger in the homozygous Pla2g6^G373R/G373R mice. The Pla2g6^KO/KO mice show a very slow progressive neurodegeneration, whereas the homozygous Pla2g6^G373R/G373R mice display an aggressive and quick neurodegenerative phenotype, yet they were both generated in a C57BL/6 background. Note that three other Pla2g6^KO/KO have been generated in C57BL/6 background and they all exhibit very similar very slow progressive phenotypes (Malik et al., 2008; Shinzawa et al., 2008; Zhao et al., 2011). Their lifespan is normal and the first rotarod assay defects are observed at ~300 days. Moreover, three other models have been described: Pla2g6^D331Y/D331Y, Pla2g6^R748W/R748W as well as a 5′ UTR transposon insertion Pla2g6^IAP/IAP (Chiu et al., 2019; Strokin et al., 2012; Sun et al., 2021). Both point mutation lines, Pla2g6^D331Y/D331Y and Pla2g6^R748W/R748W, express mutant Pla2g6 at a level comparable to their wild-type littermates (Chiu et al., 2019; Sun et al., 2021). These latter three different mouse models cause very similar and severe phenotypes when compared to the null alleles. The Pla2g6^IAP/IAP mice produce ~10% of the wild-type protein (Strokin et al., 2012). Given that four mouse models that lack the protein have much less severe phenotypes than the three models that produce Pla2g6 protein, we propose that a compensatory pathway is activated upon a complete loss of Pla2g6 gene/protein during development. Hence, to minimize genetic background issues and avoid the possible activation of a compensatory pathway, we generated and characterized transheterozygous animals (Pla2g6^KO/G373R).

To characterize the Pla2g6^KO/G373R mice, we performed rotarod and lifespan assays. As shown in Figure 2A, the control littermates (Pla2g6^+/+—Wild-type; Pla2g6^KO/+; and Pla2g6^+/G373R) show similar normal performances on rotarod assays over a period of 160 days. Homozygous Pla2g6^G373R/G373R mice show the first signs of rotarod defects between 60 and 70 days of age (Figure 2A) and die within 100 days (Figure 2B). The Pla2g6^KO/G373R mice show the first signs of rotarod defects between 80 and 90 days of age (Figure 2A) and die at ~150 days of age (Figure 2B). Hence, we decided to mainly focus on the Pla2g6^KO/G373R mice in the following experiments but used the Pla2g6^G373R/G373R mice as a comparison in some assays.

We previously showed that loss of fly homolog of PLA2G6 leads to an elevation of ceramides including GlcCer (Lin et al., 2018). GlcCer accumulation is also observed in flies that lack the fly homolog of GBA1 (Wang et al., 2022a), the gene that causes Gaucher disease (Sidransky, 2004; Wong et al., 2004). We performed immunostaining in the cerebellum and midbrain of the Pla2g6^KO/G373R and homozygous Pla2g6^G373R/G373R animals. As shown in Figure 2C, the GlcCer levels are highly upregulated in the Purkinje cells (Figure 2C; upper panel) and mid-brain cells (Figure 2C; lower panel) of both INAD mouse models when compared to controls, again showing defective ceramide metabolism. Moreover, we performed lipidomic assays to measure the levels of total Cer, total HexCer as well as Sphinganine and Sphingosine in the cerebellum of the Pla2g6^KO/G373R mice. However, unlike in flies and human cells, we did not observe an obvious change in HexCer as well as other ceramides (Figure 2—figure supplement 1D and Figure 2—source data 1). Given that the GlcCer antibody shows obvious changes that are consistent among the three species, we argue that the lipidomic assays are not sensitive enough to detect the increase in GlcCer possibly because only Purkinje cells seem to be affected.

We next explored the morphology of lysosomes in the cerebellum and midbrain of the Pla2g6^KO/G373R and homozygous Pla2g6^G373R/G373R animals. As shown in Figure 2D, we observed an expansion of LAMP1, a lysosomal marker, in the Purkinje cells in the cerebellum (Figure 2D; upper panel) and neurons of the midbrain region (Figure 2D; lower panel). Hence, lysosomal expansion is a common feature of all models of INAD.

We next performed TEM to assess the morphology of the mitochondria in the cerebellum of the homozygous Pla2g6^G373R/G373R animals. As shown in Figure 2—figure supplement 1A, the morphology of mitochondria is disrupted in Purkinje cells of the homozygous Pla2g6^G373R/G373R animals (Figure 2—figure supplement 1A–C). Moreover, we also observed a very significant increase in multivesicular bodies when compared to control animals (Figure 2—figure supplement 1B–C), consistent with an endolysosomal defect. Finally, these mice also exhibit TVSs (Figure 2—figure supplement 1B–C), a hallmark of Pla2g6 mutant animals (Sumi-Akamaru et al., 2015).

In summary, Pla2g6 mutant mice exhibit a ceramide accumulation, lysosomal enlargement as well as mitochondrial defects. These data are consistent with the fly model of INAD (Lin et al., 2018) and patient-derived NPCs (Figure 1E–H) and DA neurons (Figure 1—figure supplement 3C). Hence, we argue that these defects are evolutionary conserved and may play a critical role in the pathogenesis of INAD/PARK14.

We next tried to identify therapeutic strategies for INAD. We used INAD flies in a primary screen and then tested the drugs that improve the phenotypes in flies in INAD patient-derived cells. Based on a review of the literature, we identified 20 drugs that have been reported to control or affect sphingolipid metabolism, endolysosomal trafficking and drugs that are being tested to treat Parkinson’s disease (Figure 3—source data 2; see references in Table 4). We previously reported that INAD flies show severe bang-sensitivity at 15 days of age (Lin et al., 2018). This assay is a quick and sensitive assay that measures the propensity of flies to seize upon shock (Figure 3A). Wild-type flies right themselves in less than a second after a vortex paradigm. In contrast, bang-sensitive flies are paralyzed for an extended period of time before they can right themselves. We used this assay as our primary screen assay. As shown in Figure 3A, wild-type flies are not bang-sensitive. In contrast, INAD flies show severe bang-sensitivity and require ~20 s to recover (Figure 3A). We identified drugs that worsen the bang-sensitivity, including Fingolimod, Ozanimod, Fumonisin, Hydrochloroquine, Lonafarnib, Omigapil, and Taurine (Figure 3A). Another set of drugs did not affect bang-sensitivity, including Miglustat, Ibiglustat, NCGC607, Rapamycin, VER-155008, Deoxygalactonojirimycin, CuATSM, and Metformin (Figure 3A). However, five drugs suppressed bang-sensitivity, including Ambroxol, Genistein, PADK, ML-SA1, and Azoramide (Figure 3A). To confirm the rescue effect, we retested the drugs that suppressed the bang-sensitivity at a dosage that is 10 times higher than the dosage used in the first round. Each drug exhibited a dose-dependent increase in rescue activity, suggesting that the ability to suppress bang-sensitivity is dose-dependent (Figure 3A). To assess if these drugs modify the phenotypes in INAD patient-derived NPCs we probed Western blots of the treated cells for LAMP2. As shown in Figure 3B and C, patient-derived NPCs (29-1) do not express PLA2G6 and exhibit elevated levels of LAMP2. Correcting the variant (29-2) strongly reduces LAMP2 levels. Importantly, Ambroxol, Azoraminde, and Genistein, significantly reduce LAMP2 levels in the patient-derived NPCs (Figure 3B and C).

We previously showed that Myriocin, R55, and Desipramine reduced ceramide levels as well as the lysosomal expansion, and alleviated bang-sensitivity in the INAD fly model (Lin et al., 2018). Given that Myriocin is toxic in vertebrates and that R55 poorly penetrates the blood-brain barrier we tested Desipramine, an FDA-approved drug, in the INAD patient-derived NPCs. As shown in Figure 3B and C, Desipramine also reduces LAMP2 levels in patient-derived NPCs. We further tested whether the drugs, Ambroxol, Azoraminde, Desipramine, or Genistein, restore mitochondrial functions. As shown in Figure 3D, the ATP levels were restored upon exposure to the drugs. Taken together, these data show that the drugs not only reduce lysosomal expansion but also restore mitochondrial function.

We have previously shown that INAD flies also exhibit morphological as well as functional defects of photoreceptors (Lin et al., 2018). We therfore tested whether Ambroxol, Azoraminde, Desipramine, or Genistein can rescue these defects. As shown in Figure 3—figure supplement 1A–C, the obvious defects in electroretinograms (ERGs) profiles, including the loss of light coincident receptor potentials (LCRPs) and on-transients, are significantly improved by the drugs. In addition, the loss of photoreceptors phenotypes is also reduced in INAD flies exposed to the drugs (Figure 3—figure supplement 1D–E). In summary, we identified four drugs that suppress the loss of PLA2G6-induced phenotypes in flies and INAD patient-derived cells: Ambroxol, Azoraminde, Desipramine, and Genistein.

We previously showed that whole-body expression of human PLA2G6 in INAD flies fully rescued the neurodegenerative phenotypes and lifespan (Lin et al., 2018). In contrast, neuronal expression of human PLA2G6 strongly suppressed neurodegenerative phenotypes, but did not prolong lifespan. These data suggest that even though PLA2G6 plays an important role in the nervous system, it is also required in cells other than neurons. We, therefore, surmised that a gene therapy approach in mice should attempt the delivery of human PLA2G6 into as many cell types as possible. We, therefore, designed an AAV-based gene therapy construct, AAV-EF1a-PLA2G6 (Figure 4A). We used elongation factor EF-1 alpha (EF1a), a ubiquitous promoter, to express PLA2G6 as broadly as possible. We also designed two constructs, Lenti-CMV-PLA2G6, and AAV-EF1a-EGFP as controls (Figure 4A). The lentivirus-based construct allowed quite elevated levels of expression and was used to test toxicity when PLA2G6 is highly expressed in NPCs. The AAV-EF1a-EGFP construct was used to track viral transduction and expression efficiency. Wild-type mice injected with the AAV-EF1a-EGFP (serotype) construct via intracerebroventricular (ICV) and intravenous (IV) injections at P40 express EGFP in numerous tissues, including the cerebellum, olfactory bulb, cerebral cortex, mid-brain region, spinal cord, sciatic nerve, heart, and liver (Figure 4—figure supplement 1). However, EGFP is not expressed in the eyes or the uninjected animals (Figure 4—figure supplement 1). These data show that AAV-EF1a-EGFP is broadly expressed in many tissues when the virus is delivered via ICV and IV. However, not all cells express EGFP based on this assay.

We then conducted a pilot experiment in HEK-293T cells to assess the expression of the protein using these constructs. As shown in Figure 4B, HEK-293T expresses very low levels of endogenous PLA2G6 (upper bands marked by a *). Note that we also observed a nonspecific band (lower bands marked by #). As expected, Lenti- CMV-PLA2G6 induces very high levels of expression of human PLA2G6 in HEK-293T cells (stars in Figure 4B; last lane). However, expression of human PLA2G6 by infecting cells with AAV-EF1a-PLA2G6 packaged in either the AAV-PHP.eB or AAV9 serotype is very low, even when we used a very high multiplicity of infection (MOI of 1000) (stars in Figure 4B; lanes 3 and 4).

Upon showing that these constructs can be expressed in HEK-293T cells, we tested their expression and function in the patient-derived NPCs (Figure 4C). Expression of PLA2G6 is easily detectable in NPC cells. Moreover, the Lenti-CMV-PLA2G6 derived protein is expressed at very high levels (~9-fold higher than the endogenous levels) in the NPCs (Figure 4C; last lane). However, expression of human PLA2G6 driven by AAV-EF1a-PLA2G6 construct packaged in AAV-PHP.eB or AAV9 serotype is very low (~10% of the endogenous levels) (Figure 4C; lanes 3 and 4). Interestingly, the LAMP2 levels are reduced by about 10–20% (Figure 4C; bottom) in all three conditions, suggesting that even very low levels may have a beneficial effect.

We also examined mitochondrial morphology in these cells. As shown in Figure 4D, the patient-derived NPCs show enlarged and fragmented mitochondrial morphology (Figure 4D; upper left; arrows). Delivery of AAV-PHP.eB-PLA2G6 (Figure 4D; upper right; arrowheads) or AAV9-PLA2G6 (Figure 4D; lower left; arrowheads) strongly suppress this morphological abnormality as we observe many elongated mitochondria. In contrast, the Lenti-CMV-PLA2G6, which expresses very high levels of human PLA2G6, only partially rescues mitochondrial morphological abnormalities (Figure 4D; lower right; arrowheads). This suggests that very high levels of expression of PLA2G6 may be somewhat toxic. We also measured the ATP levels in the NPCs and found that the ATP levels are restored in the treated NPCs, arguing that mitochondrial function is at least partially restored (Figure 4E; lower panel).

In summary, both AAV-PHP.eB-PLA2G6 and AAV9-PLA2G6 induce low levels of expression of human PLA2G6 in NPCs (10–20% of the endogenous levels), and can partially alleviate lysosomal expansion as well as mitochondrial defects.

We previously showed that PLA2G6 interacts with retromer proteins, Vps26 and Vps35. This interaction is important to maintain the proper levels of Vps26 and Vps35 (Lin et al., 2018). Loss of PLA2G6 reduces the levels of Vps26 and Vps35 and impairs retromer function as well as the recycling of proteins and lipids. This leads to an accumulation of proteins and ceramides in the lysosomes and causes a lysosomal expansion. To assess whether the retromer is affected in the patient-derived cells, we assessed the levels of Vps35 in INAD patient-derived (29-1) and genetically corrected NPCs (29-2). We found that the number of Vps35 punctae is much higher in the genetically corrected NPCs (29-2) than in INAD patient-derived NPCs (29-1) (Figure 5A, a–b and B), consistent with our observations in flies that PLA2G6 increase the levels of Vps35 (Lin et al., 2018).

Next, we tested whether the selected drugs or the expression of human PLA2G6 restores Vps35 levels in INAD patient-derived NPCs. Interestingly, the addition of Ambroxol, Azoraminde, Desipramine, or Genistein to the patient-derived NPCs (29-1) did not restore the Vps35 levels (Figure 5A, a–f and B), suggesting that the drugs function downstream of the retromer. Ambroxol promotes GCase activity to degrade GlcCer in lysosomes (Magalhaes et al., 2018), Azoraminde promotes protein folding (Fu et al., 2015), Desipramine blocks the ceramide salvage pathway (Jenkins et al., 2011), and Genistein enhances lysosome biogenesis (Moskot et al., 2014). These actions are downstream of retromer function. In contrast to the drugs, expression of the human PLA2G6 restores Vps35 levels (Figure 5A, a, b, g–i and B), consistent with the model that PLA2G6 binds to Vps35 and Vps26 to maintain the levels of the retromer complex. In summary, expression of the human PLA2G6, but not the identified drugs restores the Vps35 levels.

To determine if delivery of human PLA2G6 alleviates the phenotypes in INAD mouse model, we injected five Pla2g6^KO/G373R mice with AAV-PHP.eB-EF1a-PLA2G6 construct at postnatal day 40 via ICV (4.5×10¹⁰ GC) and IV (5×10¹¹ GC) injections (Figure 6A; blue line). In an independent litter, we injected two more Pla2g6^KO/G373R mice at postnatal day 40 via ICV only (Figure 6A; purple line). Un-injected Pla2g6^KO/G373R mice displayed the first signs of rotarod impairment at ~90 days of age (Figure 6A; red line). In contrast, the ICV and IV injected Pla2g6^KO/G373R mice (Figure 6A; blue line) did not show signs of rotarod impairment until 130–140 days, a 50–60 day delay in the onset of the rotarod defects. However, even though the ICV-injected mice show some improvement in rotarod performance, statistical analyses did not show significance at most time points (Figure 6A; purple line). Note that injection of the AAV-PHP.eB-EF1a-PLA2G6 construct in wild-type littermates did not cause any obvious change in rotarod performance.

We also assessed body weight and lifespan. All injected Pla2g6^KO/G373R mice (both the ICV+IV injected group and IV injected group) exhibit a sudden body weight drop at ~190 days, 2 weeks before they die (Figure 6B). This is ~70 days later than the uninjected Pla2g6^KO/G373R animals, which show an obvious drop in body weight at ~120 days (Figure 6B). Five of the injected mice died by 210 days and two animals lived for over 300 days (Figure 6C; blue line). Figure 6 We scarified these two animals to assess the molecular changes . Hence, the injected animals live an average of at least 65 days longer than the uninjected Pla2g6^KO/G373R mice (Figure 6C; red line). In summary, our data suggest that (1) expression of human PLA2G6 in mice is safe; (2) expression of human PLA2G6 in adult pre-symptomatic Pla2g6^KO/G373R mice delays the onset of defects; and (3) promoting broad expression of PLA2G6 using ICV and IV injections is more effective than ICV delivery only.

To explore the ability of AAV-PHP.eB-PLA2G6 injections to rescue the observed defects in the cerebella of the mice, we scarified two injected Pla2g6^KO/G373R mice at ~P300. These two mice were not able to stay on the rod in a rotarod assay but behaved relatively normal in their home cage (see Video 1). As shown in Figure 7A, we found that the elevated GlcCer levels are suppressed in the P300 Purkinje cells of these two animals when compared to the uninjected Pla2g6^KO/G373R P140 mice (Figure 7A). Moreover, the levels of LAMP2 are also reduced in the Purkinje cells of these injected mice (Figure 7B). Hence, injection of the gene therapy construct restores GlcCer levels and alleviates lysosome expansion 150 days after the uninjected mice died, clearly showing beneficial effects.

Next, we assessed the levels of Vps35 in injected and uninjected Pla2g6^KO/G373R mice. The levels of Vps35 are significantly reduced in the Pla2g6^KO/G373R mice, as anticipated (Figure 7C). Interestingly, expression of human PLA2G6 not only restored the levels of Vps35, but also enhanced its expression (~2-fold of the wild-type mice) (Figure 7C). In summary, expression of human PLA2G6 using the gene therapy construct in Pla2g6^KO/G373R mice increases the Vps35 levels.

Mixed genders of flies (approximately 50% male and 50% female) were used for all experiments. Flies were raised on molasses-based food at 25°C in constant darkness. The genotypes of the flies used is y w;; iPLA2-VIA^Δ174 (Lin et al., 2018). All drugs were added freshly to regular fly food at the indicated concentration. The flies were transferred to fresh food with or without the drugs every 3 days.

To perform bang-sensitive paralytic assays, five adult flies were tested per vial. The flies were vortexed at maximum speed for 15 s and the time required for flies to stand on their feet was counted. At least 60 flies were tested per data point.

Human skin fibroblasts from Families 1 and 2 were a gift from Dr. Bénédicte Heron (Neurologie Pédiatrique Hôpital Trousseau) and Dr. Young-Hui Jiang (Duke University), respectively. The control skin fibroblasts (Cat. #GM23815) were purchased from Coriell Institute.

Human iPSCs (hiPSCs) maintained on Cultrex (CTX Cat# 3434-010-02, R&D Systems) in PSC Freedom Media (FRD1, ThermoFisher Custom) were passaged every 4–5 days using Stem-Accutase (Cat# A11105-01, Life Technologies) in the presence of 1 µM Thiazovivin (THZ Cat# SML1045-25 MG, Sigma-Aldrich).

Undifferentiated iPSCs were grown following standard protocols in StemFlex medium on laminin521 substrate (BioLamina) as described previously (Ruzo et al., 2018). iPSCs were differentiated to ventral midbrain NPCs and DA neurons following (Kim et al., 2021) with the following modifications. Neural induction was initiated at day 0 by passaging iPSC, single cell seeding at 1 M cells/ml in 2.5 M/well of over ultralow adherence plastic 6wp (Corning, Cat #3471vendor) in neural induction medium (base medium changed to AaDMEMEMM/F12:Neurobasalbobasla 50:50, with N2, B27 without RA, RhoK inhibitor Y27632 10 µM (first 2 days), 100 nM LDNn, 10 µM SB-431542 and, substituting 1 mM SAG3.1 and 1 mM Prurmorprphhpamione (R&D) in place of SHH protein), and 1 µM CHIR99021 (R&D). At days 4–7, CHIR99021 was increased to 6 µM, and from days 8 to 11 was decreased to 3 µM and 0.2 mM AA (Sigma-Aldrich), 0.2 mM dbcamp (Sigma-Aldrich), 10 µM DAPT (Tocris), and 1 ng/ml TGFb3 (R&D) 0 were added from day 10 onward. EBs were dissociated with accutase at day 16, and reseeded at 0.8Mcells/cm² on polyornithine laminin (1 mg/ml in pH 8.4 borate buffer followed by 10 µg/ml natural mouse laminin in DMEM/F12) coated TC plastic for continued differentiation and day 22 fixation and staining for positional markers, or frozen in 2× FM (Millipore Sigma; ES-002-10F). Live cultures or thawed cells were either expanded as floor plate progenitors (following Brundin approach; Floor Plate Progenitor Kit expansion protocol; Thermo Fisher Scientific; A3165801) or terminally differentiated to DA neurons with all prior factors except in neurobasal base medium after day 25.

Cells were homogenized in 1% NP40 lysis buffer (20 mM HEPES pH7.5, 150 mM NaCl, 1% NP-40, 10% Glycerol, and Roche protease inhibitor mix) on ice. Tissue or cell debris were removed by centrifugation. Isolated lysates were loaded into 10% gels, separated by SDS-PAGE, and transferred to nitrocellulose membranes (Bio-Rad). Primary antibodies used in this study were as follows: mouse anti-PLA2G6 antibody (Santa Cruz Biotechnology; sc376563), rabbit anti-PLA2G6 antibody (Sigma-Aldrich; SAB4200129), rabbit anti-PLA2G6 antibody (Sigma-Aldrich; HPA001171), mouse anti-Actin (ICN691001, Thermo Fisher Scientific), and rabbit anti-LAMP2 antibody (Abcam; ab18528).

The cultured cells were fixed with 4% paraformaldehyde (PFA) in 1× PBS at 4°C for 30 min. The fixed cells were permeabilized in 0.1% Triton X-100 in 1× PBS for at least 15 min at room temperature. For mouse tissues, the mouse was deeply anesthetized using isoflurane, and perfused intracardially with 1× PBS and 4% PFA. Brains were dissected and post-fixed in 4% PFA/PBS overnight at 4°C. The next day, tissues were cryoprotected in a 20% sucrose/PBS solution at 4°C for 1 day, followed by a 30% sucrose/PBS solution at 4°C for one more day. Tissues were then embedded and frozen in OCT and cryosectioned using a cryostat (Leica CM1860) at 25–40 μm. The fixed mouse tissues were permeabilized in 0.1% Triton X-100 in 1× PBS for at least 15 min at room temperature. The following antibodies were used for the immunofluorescence staining: chicken anti-Tyrosine Hydroxylase antibody (Abcam; ab76442; RRID:AB_1524535); rabbit anti-LAMP2 antibody (Abcam; ab18528; RRID:AB_775981); mouse anti-TOMM20 antibody (Abcam; ab56783; RRID:AB_945896); rabbit anti-GlcCer (Glycobiotech; RAS_0010); mouse anti-ATP5a (Abcam; ab14748; RRID:AB_301447); mouse anti-LAMP1 (Abcam; ab25630; RRID:AB_470708); goat anti-Vps35 (Abcam; ab10099; RRID: AB_296841); mouse anti-GFP FITC conjugated (Santa Cruz Biotechnology; sc9996; RRID:AB_627695); NESTIN (Millipore; ABD69; RRID:AB_2744681); FOXA2 (R&D; AF2400; RRID:AB_2294104); LMX1A (Millipore; AB10533; RRID:AB_10805970); OTX2 (R&D; AF1979; RRID:AB_2157172); FOXG1 (Takara; ABM227); and Alexa 488-, Cy3-, or Cy5-conjugated secondary antibodies (111-545-144, 111-585-003, and 111-175-144); Jackson ImmunoResearch Labs; or Alexa 488, 555, 647 conjugated anti-rabbit, mouse, and goat (Thermo Fisher Scientific). All the confocal images were acquired with a Model SP8 confocal microscope (Leica), with the exception of vmNPC characterization and initial Da neuron characterization, acquired with a Phenix spinning disk confocal HTS microscope (PE) and LSM700 Confocal microscope (Zeiss), respectively. Confocal images were processed using ImageJ and Photoshop (Adobe).

Construction of pAAV-EF1a-PLA2G6: PCR amplified fragments of the full-length human PLA2G6 cDNA (NM_001349864.2) containing XbaI (5′) and EcoRV (3′) were digested with XbaI and EcoRV. To generate the vector backbone, pAAV-EF1a-CVS-G-WPRE-pGHpA (Addgene; 67528) was digested with XbaI and EcoRV and gel eluted to remove the CVS-G. The digested PCR fragments were ligated in the digested vector backbone to generate pAAV-EF1a-PLA2G6-WPRE-pGHpA. The pGHpA was replaced by SV40pA to reduce the total size of the construct. Similarly, the pAAV-EF1a-EGFP was constructed following the same method except the EGFP cDNA was used. Lenti-CMV-PLA2G6 was purchased from Horizon Discovery-Precision LentiORF collection (OHS5897-202617053).

A plasmid DNA cocktail solution was produced by combining the pAAV transgene (pAAV-EF1a-PLA2G6 or pAAV-EF1a-EGFP), rep/cap serotype PHP.eB, and Ad∆F6 helper plasmids and transfected into HEK293T cells with iMFectin poly transfection reagent – GenDEPOT. A total of 80×15 cm² dishes were overlaid with the DNA cocktail solution and were allowed to incubate for 4 hr before adding fresh media. Three days post-transfection, dishes were imaged, harvested, and digested. The collected cell viral lysate was centrifuged, supernatant was transferred to a new tube and digested separately. The cell pellet was resuspended in TMN (Tris-HCl cell suspension buffer) and cell-associated AAV was recovered by cell lysis treatment with 5% sodium deoxycholate, DNase, RNase, and three subsequent freeze/thaw cycles. Media-secreted AAV was precipitated in a 40% polyethylene glycol solution. Digest CVL was centrifuged to remove cell debris and cleared supernatant was transferred to a new tube. PEG precipitated supernatant was centrifuged, the resulting pellet was resuspended in HBS (HEPES Buffered Saline) solution. After digestion and precipitation, cleared cell viral lysate and media secreted AAV were combined and purified on a discontinuous iodixanol gradient. The band corresponding to purified AAV was extracted and diluted in DPBS supplemented with 0.001% Pluronic F-68. Viral diluent was then concentrated in Amicon centrifugation filtration units (100,000 MW) to the desired volume. Concentrated AAV was first diluted 1:100 and then serially diluted 10-fold to yield AAV dilutions of 0.01, 0.001, 0.0001, and 0.0001. The titer of the AAV vectors was quantified with the primers corresponding to WPRE and probed against a GVC in-house standard. The AAV-PHP.eB was packaged by the Gene Vector Core at Baylor College of Medicine.

All experimental animals were treated in compliance with the United States Department of Health and Human Services and the Baylor College of Medicine IACUC guidelines. Mice were reared in 12-hr light-dark cycles with access to food and water ad libitum. The mouse lines used here include the Pla2g6 complete knockout line (genotype: Pla2g6^KO/KO) (Bao et al., 2004) and the Pla2g6^G373R point mutation mouse line (Genotype: PLA2G6^G373R/G373R) (Wada et al., 2009). The PLA2G6 complete knockout line was a donation from Dr. Sasanka Ramanadham at the University of Alabama at Birmingham. Cryopreserved sperm of the Pla2g6^G373R mice (No. BRRC04196) was purchased from RIKEN BioResource Research Center. The cryopreserved sperms were used to in vitro fertilize female C57BL/6 mice to retrieve the line. The in vitro fertilization was performed by the GERM Core at Baylor College of Medicine. After the Pla2g6^G373R point mutation mice were retrieved, we mated P la2g6^KO/+ with P la2g6^G373R/+ mouse to generate transheterozygous mice (Genotype: P la2g6^KO/G373R). Mice were genotyped by standard PCR using isolated tail DNA as template.

Preoperatively, adult mice were weighed and given meloxicam 30 min prior to surgery with a 30-gauge needle to minimize discomfort. Animals were anesthetized preoperatively with isoflurane. Following anesthesia, animals were checked for pedal withdrawal reflex; if absent, the animals were transferred to stereotaxic apparatus and maintained under anesthesia using volatilized isoflurane (1–3% depending on physiological state of the animal, which was continuously monitored by response to tail/toe pinch). Isoflurane was diluted with pressurized oxygen using an anesthetic vaporizer system that allows precise adjustment of isoflurane and gas flow to the animal. Surgical site was prepared by shaving hair, administering depilatory cream and three times betadine surgical scrub, three times ethanol wipe. A short incision (0.5–2 cm) was made over the skull using small surgical scissors. Following the incision, a small burr hole craniotomy (<1 mm in diameter) was made using a dental drill 0.5 µl of AAV particles were injected into the third ventricle (coordinates from Bregma in mm: ML 1, AP –0.3, DV −2.3) using a 33-gauge Hamilton syringe. Post-injection, the scalp was sutured using surgical nylon monofilament (Ethicon Cat# 1689G) in a simple interrupted pattern for skin-to-skin closure. After 7 days, mice were briefly anesthetized with isoflurane and sutures were gently removed using small scissors and forceps.

Random selected mice were placed in restrainer to immobilize and have easy access to the tail. Tail was soaked in warm water for 10–15 s to cause vasodilation (enlargement) of the vein and was swabbed with a gauze dampened with alcohol to increase the visibility of the vein. One of the two lateral tail veins was located in the middle third of the tail. With the bevel of the needle facing upward and the needle almost parallel to the vein, the needle was slid into the tail vein. By gently applying negative pressure to the plunger and observing a flash of blood in the hub vein penetration was confirmed. Slowly the plunger was pressed to inject AAV-containing solution into the vein. The needle was removed from the vein and slight pressure was applied to the puncture site with a dry piece of gauze until the bleeding has stopped. The animal was removed from its restrainer and placed in the cage. The animal was monitored for 5–10 min to ensure hemostasis and that the bleeding has stopped.

We measured the time (latency) a mouse takes to fall off the rod under continuous acceleration. On the day of testing, mice were kept in their home cages and acclimated to the testing room for at least 15 min (acclimation phase). The apparatus was set to accelerating mode from 5 to 40 rpm in 300 s. We record the latency at which each mouse falls off the rod. The first trial is the training period. During the training period, the mice were tested three times separated by 15 min intervals in a day and the trials were repeated for 3 consecutive days. After the training period, the mice were tested once a week. At the end of the rotarod assay, we weighted each mouse every other week. All the mice were tested in blind of genotype and treatments.

For fly experiments, sample sizes are stated in the figure legends. The drug treatments were performed for three biological replicates. For all cell experiments, the studies were conducted in parallel with vehicle controls in the neighboring well for at least three biological replicates. All mouse experiments were conducted in the number of technical replicates indicated in the figure legends. Error bars are shown as standard error of the mean (SEM). The mouse survival data sets (Figure 2B and Figure 5C) were organized and analyzed in GraphPad Prism 9.4.0 using Mantel-Cox and Gehan-Breslow-Wilcoxon tests. All other data sets from flies and cells were organized and analyzed in Microsoft excel 2010 using two-tailed Student’s t test. The criteria for significance are NS: not significant; p>0.05; *p<0.05; **p<0.01, and ***p<0.001.

Lipids were extracted from the method as described previously from fibroblast cells and cerebellum tissue samples (PMID 26090945, 35857503). d18:1/16:0-d3-GlcCer, deuterated ceramide LIPIDOMIX (mixture of d18:1-d7/16:0-, d18:1-d7/18:0-, d18:1-d7/24:0-, and d18:1-d7/24:1-Cer), d18:1-d7 sphingosine (So), and d18:0-d7-sphinganine (Sa), all purchased from Avanti Polar Lipids, Inc, were used as internal standards (IS).

For fibroblasts, IS mixture was added equally to cell pellets equivalent to 1×10⁶ cells with 500 µl 1-butanol:methanol (1:1, v/v) as extraction solvent. Samples were tip-sonicated on ice, followed by water bath sonication for 60 min at room temperature and centrifugation at 13,000×g for 10 min. The supernatant was collected and dried using nitrogen air and reconstituted in chloroform:methanol (1:4, v/v) and centrifuged at 13,000×g for 5 min to remove any insoluble salt.

For tissue samples, all tissue samples were hard-frozen in liquid nitrogen, followed by cryopulverization. Powdered samples were sonicated in water bath for 60 min and centrifuged at 13,000×g for 10 min. The supernatant was collected and dried using nitrogen air and reconstituted in chloroform:methanol (1:4, v/v) and centrifuged at 13,000×g for 5 min to remove any insoluble salt.

Vanquish Horizon UHPLC (Thermo Fisher Scientific, Waltham, MA) was coupled to Thermo Altis triple quadrupole (Thermo Fisher Scientific, Waltham, MA) for targeted analysis of lipids. Sphingosine, sphinganine, hexosylsphingosine, hexosylceramides, and ceramides were analyzed on a Atlantis T3 (1.0×50 mm², 3 µm) under a binary gradient of mobile phase A (water:acetonitrile, 9:1, v/v) and mobile phase B (isopropanol:methanol:acetonitrile, 7:2:1, v/v/v) with 0.1% formic acid at the flow rate of 150 µl/min. Mobile phase B was increased from 10% to 30% over 1 min, 40% over 1 min, 70% over 1.5 min, 72.5% over 3.5 min, and 95% over 3.5 min. After maintaining the mobile phase B at 95% for 1.5 min, it was ramped down to 10% over 0.1 min and re-equilibrated at 10% mobile phase B for 3 min. Scheduled multiple reaction monitoring of targeted lipids were acquired with Q1 resolution at 0.4 FWHM, Q3 resolution at 0.7 FWHM, spray voltage 3.5 kV, sheath gas at 35, auxiliary gas at 5, and sweep gas at 1 in positive ion mode. Quantifier ion of m/z 264.25 and qualifier ion of m/z 282.25 were used for Cer species containing d18:1 long-chain base. For Cer species containing d18:0 long-chain base, m/z 284.25 as quantifier ion and m/z 266.25 as qualifier ion were used. For HexCer species, m/z 264.25 as quantifier ion and m/z equivalent to loss of hexose ring from the precursor ion was used as qualifier ion. Product ions of m/z 282.25 for So and hexosylsphingosine, and m/z 284.25 for Sa were utilized as quantifier ion. The analytical column was kept at 40°C throughout the analysis.

All lipids were quantified by calculating the peak areas and ratio of analyzed lipids over the IS’s peak areas were calculated. Ratio of lipids against IS were normalized by the protein amount measured the samples using BCA protein assay. The total level of ceramides and hexosylceramides were calculated by adding the normalized peak areas ratios of individual species within each class.