Lysosomal TPC2 channels disrupt Ca²⁺ entry and dopaminergic function in models of LRRK2-Parkinson’s disease

Parkinson’s disease (PD) is the most common movement disorder and the second most common neurodegenerative disease after Alzheimer’s disease (Balestrino and Schapira, 2020). The dominant early motor features of PD result from the selective degeneration of dopaminergic neurons within the substantia nigra pars compacta. Additional nonmotor symptoms such as cognitive dysfunction are increasingly common as the disease progresses (Schapira et al., 2017) with others such as visual changes even manifesting during the presymptomatic prodrome to PD (McNeill et al., 2013). Mutations in the leucine-rich repeat kinase 2 (LRRK2) gene are a common cause of familial PD, and patients carrying them present with a phenotype clinically and pathologically indistinguishable from idiopathic disease (Di Fonzo et al., 2006; Paisán-Ruíz et al., 2004; Zimprich et al., 2004). Pathogenic LRRK2 drives deficits in neuronal morphology (MacLeod et al., 2006) and synaptic transmission (Afsari et al., 2014; Arranz et al., 2015; Matta et al., 2012; Piccoli et al., 2014) including dopamine and glutamate release (Liu et al., 2015; Volta et al., 2017). LRRK2 is a large, multidomain signaling protein with the most common G2019S PD-linked mutation falling within the kinase domain (West et al., 2005). A subset of the Rab family of trafficking GTPases are bona fide LRRK2 substrates (Steger et al., 2016), but few others have been confirmed. Thus, the function of LRRK2 remains enigmatic with a clear need to delineate downstream targets as alternative ways to combat LRRK2 defects and PD more generally.

Ca²⁺ is key for neuronal homeostasis (Berridge, 1998). Oscillatory Ca²⁺ influx into the cytoplasm is associated intimately with spontaneous firing of dopaminergic neurons in the substantia nigra (Guzman et al., 2009; Nedergaard et al., 1993). This flux through voltage-gated Ca²⁺ channels has attracted therapeutic attention not least due to epidemiological evidence linking Ca²⁺ channel blocker use to reduced risk of PD (Chan et al., 2007; Ritz et al., 2010). Downstream of Ca²⁺ entry, excessive Ca²⁺ uptake by mitochondria potentially links Ca²⁺ homeostasis disruption to the established role of mitochondria and oxidative stress in the disease (Guzman et al., 2010; Surmeier et al., 2017). In addition to mitochondrial dysfunction, growing evidence links lysosomal dysfunction to PD (Dehay et al., 2013; Platt, 2014; Smith et al., 2022). Genetics have identified mutations in the GBA1 gene, which encodes a lysosomal hydrolase (Sidransky et al., 2009), as the most common risk factor for PD. Additionally, large-scale sequencing efforts of sporadic PD patients identified an excessive burden of variants in genes that cause lysosomal storage diseases (Robak et al., 2017). LRRK2 localizes to the endo-lysosomal system (Alegre-Abarrategui et al., 2009; Biskup et al., 2006; Dodson et al., 2012; Schapansky et al., 2018) particularly upon stress (Eguchi et al., 2018) and has been heavily implicated in endo-lysosomal trafficking (Madureira et al., 2020; Manzoni and Lewis, 2013), morphology, and repair (Bonet-Ponce et al., 2020; Herbst et al., 2020). There is much evidence now that the endo-lysosomal system serves as a functionally relevant Ca²⁺ store (Galione and Muallem, 2023; Patel and Muallem, 2011). But the potential role of lysosomal Ca²⁺ signaling dysfunction in PD and neurodegeneration more generally is underexplored (Patel, 2016).

Here, we used human and fly models, a number of in vitro and in vivo assays including Ca²⁺ imaging together with novel pharmacological manipulation to identify a role for the lysosomal cation channel two-pore channel-2 (TPC2) in Ca²⁺ dysfunction and dopaminergic deficiency in LRRK2 PD. TPC2, together with TPC1, is an ancient member of the voltage-gated ion channel superfamily regulated by the signaling molecules, NAADP and PI(3,5P)₂ (Brailoiu et al., 2009a; Calcraft et al., 2009; Wang et al., 2012). TPC2 has the highly unusual ability to function as a Ca²⁺ or Na⁺ channel depending on the activating stimulus (Gerndt et al., 2020; Yuan et al., 2022). It is a Rab effector (Abrahamian et al., 2024; Lin-Moshier et al., 2014), regulates numerous cellular processes including many aspects of membrane traffic (Grimm et al., 2014; Kilpatrick et al., 2017; Sakurai et al., 2015), and has previously been shown to mediate autophagic dysfunction (Gómez-Suaga et al., 2012), lysosomal morphology defects (Hockey et al., 2015), and aberrant activation of the lysosomal transcription factor, TFEB (Nabar et al., 2022), in response to mutant LRRK2 in nonexcitable cells (Hockey et al., 2015). Here, we report neuronal Ca²⁺ entry defects in response to pathogenic LRRK2 in vitro, a key role of TPC2 in mediating these defects and TPC2-dependent disruptions in dopaminergic circuits that drive vision and movement in vivo. We further exploit the malleable ion selectivity of TPC2 to pharmacologically restore Ca²⁺ homeostasis and behavioral well-being. These data suggest a novel strategy for slowing PD.

To examine the effects of pathogenic LRRK2 on neuronal Ca²⁺ homeostasis, we generated dopaminergic SH-SY5Y cell lines stably expressing either wild-type LRRK2 or the common PD-causing G2019S mutant (Fig. 1 a). We analyzed expression by western blotting (Fig. S1 a) and quantitative PCR (Fig. S1 b) and selected two clones of each genotype with differing expression levels for our analyses. LRRK2 expression was lower than in cell lines established previously (Papkovskaia et al., 2012) and much nearer endogenous levels (Fig. S1 a), thereby providing a more physiological readout of LRRK2 action.

To measure neuronal activity, we monitored cytosolic Ca²⁺ levels in individual differentiated cells loaded with the Ca²⁺ indicator Fura-2 (Fig. 1 b). Cells were depolarized with 50 mM K⁺ to stimulate Ca²⁺ entry from outside of the cell via voltage-gated Ca²⁺ channels. As shown in Fig. 1 c, depolarization induced Ca²⁺ transients in the parental line. The responses were similar in cells expressing the empty vector. In marked contrast, Ca²⁺ responses in the two independent cell lines expressing LRRK2 G2019S were increased (Fig. 1 d). To assess the specificity of this effect, we took two approaches. In the first approach, we measured depolarization-induced responses in cells expressing wild-type LRRK2. Ca²⁺ entry was reduced relative to the mutant lines in both of the lines tested resulting in signals comparable to the cells lacking exogenous LRRK2 (Fig. 1 e). In the second approach, we analyzed cell lines expressing kinase-inactive LRRK2 (Fig. S1 c). Again, Ca²⁺ entry was reduced in both of the lines tested relative to the LRRK2 G2019S lines and similar to the control cells (Fig. S1, d and e). We also analyzed lines expressing the R1441C or R1441G mutations within the ROC/COR domain of LRRK2 (Fig. S1 f). These lines showed exaggerated Ca²⁺ signals although the effects were less pronounced than the G2019S mutation (Fig. S1, g and h). Summary data quantifying both the peak response and the area under the curve across lines are shown in Fig. 1, f and g; and Fig. S1 i.

We used automated plate reading coupled with microfluidics to obtain full concentration–effect relationships for extracellular K⁺ and intracellular Ca²⁺ (Fig. 1 h). K⁺ evoked Ca²⁺ responses in a concentration-dependent manner in cells expressing the empty vector (Fig. 1 i). Ca²⁺ responses in cells expressing LRRK2 G2019S were increased at all K⁺ concentrations relative to the empty vector. In contrast, Ca²⁺ responses in cells expressing wild-type LRRK2 were not (Fig. 1 i).

We also performed experiments using cells expressing the genetically encoded Ca²⁺ indicator, GECO1.2 (Fig. 1 b). This was to mitigate against any off-target effects of the chemical indicator for Ca²⁺ on membrane potential (Smith et al., 2018). As shown in Fig. 1 j, GECO1.2 expression was readily detectable in wild-type and LRRK2 G2019S–expressing cells. Robust depolarization-evoked Ca²⁺ responses were recorded, and similar to the responses in cells loaded with Fura-2, the responses were larger in the LRRK2 G2019S line (Fig. 1 k). In contrast, Ca²⁺ signals evoked following subsequent stimulation of the same cells with the Ca²⁺ ionophore ionomycin were similar (Fig. 1, k and l), attesting to specificity.

To further test the specificity of the G2019S effect, we compared store-operated Ca²⁺ entry in the wild-type and mutant cells. Endoplasmic reticulum Ca²⁺ store depletion was induced by treating cells with the SERCA inhibitor thapsigargin in the absence of external Ca²⁺ (Fig. 1 m) As shown in Fig. 1 n, thapsigargin evoked a Ca²⁺ signal consistent with leak of Ca²⁺ from the endoplasmic reticulum. Subsequent addition of external Ca²⁺ evoked Ca²⁺ entry. Neither the initial response to thapsigargin nor the subsequent one to Ca²⁺ was different in the LRRK2 G2019S cells compared with the controls (Fig. 1, n and o), again attesting to specificity.

In sum, we show depolarization-induced Ca²⁺ entry is selectively exaggerated by pathogenic LRRK2.

Lysosomes are increasingly implicated in the actions of LRRK2 and PD more generally (Dehay et al., 2013; Platt, 2014; Smith et al., 2022). To probe the role of lysosomes in Ca²⁺ dysfunction mediated by LRRK2 G2019S, we began by disrupting lysosome integrity. We did this in two ways (Fig. 2 a). In the first, we treated cells with the lysosome-permeabilizing agent LLOMe (Fig. 2 a). As shown in Fig. 2 b, LLOMe changed the subcellular distribution of endocytosed dextran from a punctate one to a more diffuse one consistent with a mild permeabilizing effect. This treatment was sufficient to reverse the potentiating effects of the G2019S mutant on depolarization-induced Ca²⁺ entry (Fig. 2 e). Essentially, similar results were obtained with a structurally distinct cathepsin C substrate, GPN (Fig. S2, a and b). This was not due to changes in lysosomal pH (Atakpa et al., 2019) because NH₄Cl did not affect Ca²⁺ entry in LRRK2 G2019S–expressing cells (Fig. S2, a and b). In the second, we treated cells with vacuolin-1, which promotes endo-lysosomal fusion (Fig. 2 c). Consistent with this action, vacuolin-1 induced the appearance of large cytoplasmic vesicles (Fig. 2 d). As with LLOMe, vacuolin-1 also reversed Ca²⁺ defects induced by the G2019S mutant (Fig. 2 e). To probe the role of the endoplasmic reticulum in LRRK2 G2019S action, we depolarized cells following depletion of endoplasmic reticulum Ca²⁺ stores with thapsigargin. Thapsigargin had little effect on depolarization-induced Ca²⁺ entry (Fig. 2 f). Taken together, these data summarized in Fig. 2 g show that interfering with lysosomal integrity reverses Ca²⁺ defects evoked by LRRK2 G2019S.

To probe the mechanisms underlying LRRK2-mediated Ca²⁺ defects, we considered a role for TPC2 because TPC2 is a lysosomal ion channel and inhibiting it reverses trafficking defects in LRRK2 G2019S patient fibroblasts (Hockey et al., 2015). Ca²⁺ release through TPC2 is activated indirectly by the second messenger NAADP (Fig. 2 h) (Marchant et al., 2022). We targeted this axis in two ways. In the first pharmacological approach, we used the recently described NAADP antagonist PF-543 (Gunaratne et al., 2018). As shown in Fig. 2 i, PF543 reversed the effects of LRRK2 G2019S. We also targeted the channel more directly with the TPC2 blocker tetrandrine (Fig. 2 g). Tetrandrine reversed the effects of LRRK2 G2019S (Fig. 2 i). In contrast, chemical inhibition of lysosomal TRP mucolipin channels with ML-SI3 had little effect (Fig. S2, c–e). In the second approach, we expressed TPC2 mutated at Leu265 within the pore region (Fig. 2 h) (Brailoiu et al., 2010b). This construct (TPC2^PD, for pore-dead) acts in a dominant negative way to block NAADP action. GFP-tagged TPC2^PD was expressed on punctate structures similar to LAMP1-GFP (Fig. 2 j) and colocalized with endogenous LAMP1 similar to wild-type TPC2 (Fig. S2 f). As shown in Fig. 2 k, depolarization-evoked Ca²⁺ entry in LRRK2 G2019S–expressing cells was reduced in cells expressing TPC2^PD relative to neighboring untransfected cells. In contrast, the overexpression of LAMP1 had little effect. Pooled data quantifying the effects of PF-543, tetrandrine, and TPC2^PD on Ca²⁺ entry are shown in Fig. 2 l.

In sum, both chemical and molecular strategies inhibiting TPC2 normalized defective Ca²⁺ entry caused by pathogenic LRRK2.

To investigate the link between lysosomal Ca²⁺ release and Ca²⁺ entry, we leveraged the availability of recently described cell-permeable TPC2 agonists that bias the channel to either a Ca²⁺-permeable, NAADP activated–like state or a more Na⁺-selective, PI(3,5)P₂ activated–like state (Fig. 3 a) (Gerndt et al., 2020).

As shown in Fig. 3 b, the NAADP-mimetic TPC2-A1-N evoked a modest but detectable Ca²⁺ signal in parental SH-SY5Y cells. These experiments were performed in the absence of external Ca²⁺ to isolate Ca²⁺ release. TPC2 is also activated by PI(3,5)P₂, but in stark contrast to NAADP, it evokes largely Na⁺-selective currents (Gerndt et al., 2020). This “switch” can be mimicked by TPC2-A1-P (Fig. 3 a). TPC2-A1-P had little effect on cytosolic Ca²⁺ consistent with signaling through Na⁺ (Fig. 3 b). Costimulation of TPC2 with its agonists, however, caused a significantly increased Ca²⁺ response (Fig. 3, b and c), indicating that the recently reported synergistic activation of TPC2 in nonexcitable cells (Yuan et al., 2022) is a feature of excitable cells too.

We also examined TPC2 activation in the presence of external Ca²⁺ (Fig. 3 d). Strikingly, the agonist combination evoked a robust Ca²⁺ signal comparable in amplitude to that evoked by depolarization. This effect was specific as external Ca²⁺ only modestly increased the effects of TPC2-A1-P (Fig. 3, c and d). Activation of TPC2 evoked similarly robust Ca²⁺ responses in cells expressing wild-type and G2019S LRRK2 (Fig. 3 f). These effects were concentration-dependent (Fig. S3). To further test specificity, we used the inactive TPC2-A1-N analog, SGA-10 (Fig. 3 a). SGA-10 in combination with TPC2-A1-P had only modest effects on cytosolic Ca²⁺ (Fig. 3, e and f).

Taken together, these data reveal that TPC2-evoked Ca²⁺ release from lysosomes is coupled with Ca²⁺ entry, consistent with a requirement for TPC2 in deviant depolarization-induced Ca²⁺ entry mediated by LRRK2 G2019S.

Activation of TPC2 with TPC2-A1-N increases lysosomal pH and decreases lysosomal motility, whereas channel activation with TPC2-A1-P increases lysosomal exocytosis (Gerndt et al., 2020). Lysosomal exocytosis may represent a route to clear defective lysosomes and their contents. We therefore examined the effect of TPC2-A1-P on depolarization-induced Ca²⁺ signals. TPC2-A1-P was without effect on cytosolic Ca²⁺ levels in LRRK2 G2019S–expressing cells (Fig. 4 a). But, strikingly, TPC2-A1-P reversed the deviant effects of LRRK2 G2019S on depolarization-evoked Ca²⁺ entry (Fig. 4 a). We also tested riluzole, a structurally distinct TPC2 activator with a biophysical current profile similar to TPC2-A1-P (Fig. 3 a) (Zhang et al., 2019). Like TPC2-A1-P, riluzole also reset the Ca²⁺ signals (Fig. 4 a). These data, summarized in Fig. 4 b, show that select activation of TPC2, as well as inhibition (Fig. 3), is corrective.

Thus far, all experiments were performed in neuroblastoma cells overexpressing LRRK2 (albeit modestly). This model might not recapitulate the situation in diseased neurons. To address this, we generated human midbrain dopaminergic neurons differentiated from PD patient–derived induced pluripotent stem cells (iPSCs; Fig. 4, c and d). These lines are heterozygous for the G2019S mutation. As shown in Fig. 4, e and f, TPC2 agonists acted synergistically to evoke robust Ca²⁺ signals in these neurons just as they did in SH-SY5Y cells. Similar results were obtained in two independent PD lines (Fig. 4 f; and Fig. S4, a and b). Depolarization induced Ca²⁺ signals in both healthy controls and PD lines (Fig. 4 g) albeit variably (Fig. S4, c–f). Importantly, Ca²⁺ signals in PD lines were reduced by TPC2-A1-P (Fig. 4, g and h) just as they were in SH-SY5Y cells, in all three lines tested (Fig. S4, g and h). This analysis extends our findings to human neurons expressing LRRK2 at endogenous levels and validates findings in neuroblastoma cells.

In sum, deviant Ca²⁺ entry evoked by pathogenic LRRK2 can be rescued by biasing cation flux through TPC2 in two independent in vitro models.

We next tested whether TPC2 in itself was sufficient to disrupt dopaminergic neuron function. For these experiments, we leveraged the genetic tractability of Drosophila in which the G2019S variant of human LRRK2 or its fly equivalent perturbs visual and motor function (Afsari et al., 2014; Cording et al., 2017; Fellgett et al., 2021). As shown in Fig. 5 a, orthologs of the TPC gene (TPCN) were not readily identifiable in Drosophila and other Dipterans. But they were present in a number of insect orders including the closely related Lepidoptera and Siphonaptera. Synteny analyses showed that the TPC gene when present was in general flanked by MVK and RPL6 (Fig. 5 b). These genes are located on different chromosomes in Diptera, indicating that loss of the TPC gene was likely associated with chromosomal rearrangements. Consistent with this, ALG and BOLA1 that are normally upstream neighbors are also located on different chromosomes in Diptera. Thus, TPCs have undergone lineage-specific localized loss. This renders Drosophila an ideal null background for assessing human TPC2 functionality.

We first expressed human TPC2 tagged with mCherry in Drosophila S2R⁺ cells and examined its localization by confocal microscopy. As shown in Fig. 5 c, TPC2 localized to vesicular structures similar to the lysosomal marker LAMP1. The coexpression of TPC2 and LAMP1 confirmed colocalization of the two (Fig. 5, d and e). Similar results were obtained with the pore-dead mutant, TPC2^PD (Fig. 5, d and e).

We also expressed TPCs tagged with the genetically encoded Ca²⁺ indicator GCaMP6s (Gerndt et al., 2020) to monitor channel activity (Fig. 5 f). As with the mCherry-tagged constructs, TPC2- and TPC2^PD-GCaMP6s localized to vesicular structures (Fig. 5 c). Stimulation of TPC2-GCaMP6s with a combination of TPC2-A1-N and TPC2-A1-P resulted in a Ca²⁺ signal, albeit modest relative to ionomycin (Fig. 5 g). This signal was specific because the TPC2-A1-N analog SGA-10 combined with TPC2-A1-P had little effect on GCaMP6s fluorescence (Fig. 5 g). Additionally, as shown in Fig. 5 g, the TPC2 agonist combination did not evoke a response in cells expressing TPC2^PD-GCaMP6s, but ionomycin responses were similar. These data are summarized in Fig. 5 h.

To further test handling of human TPC2 by Drosophila cells, we disrupted the endo-lysosomal targeting motif (Leu11/Leu12) in TPC2 (Fig. 5 i). The expression of TPC2^PM (for plasma membrane) tagged with mCherry in S2R+ cells resulted in peripheral and intracellular labeling (Fig. 5 d). Colocalization analyses with a plasma membrane marker confirmed rerouting away from the lysosomes (Fig. 5 e). Ca²⁺ imaging experiments using cells labeled with Fura-2 revealed that rerouted TPC2 responded to the TPC2-A1-N and TPC2-A1-P combination. Again, this signal was specific because SGA-10 when combined with TPC2-A1-N was without effect. And there was little influx in cells expressing the empty vector (Fig. 5, j and k).

Taken together, we show that heterologously expressed human TPC2 traffics and functions normally in Drosophila cells, which lack TPC orthologs.

Having established the functionality of human TPC2 in Drosophila in vitro, we generated transgenic flies expressing human TPC2 in dopaminergic neurons under the tyrosine hydroxylase promoter. We used the bipartite GAL4-UAS system to insert transgenes in the AttP2 genomic landing site (Fig. 6 a) (Bischof et al., 2007). These animals were analyzed in parallel with those expressing LRRK2 G2019S (Fig. 6 a). As shown in the confocal images in Fig. 6 b, TPC2 was readily detectable in tyrosine hydroxylase–positive neurons. This included staining of PPL2 neurons, which control vision, and TH-VUM, which controls movement. To examine the functional consequences of TPC2 expression in intact flies, we analyzed visual responses in the retina and locomotion (Fig. 6 c).

Vision defects are a common nonmotor symptom of PD (Schapira et al., 2017). Visual responses are controlled by dopamine neurons, which ramify the medulla and optic lobe. We used steady-state visual evoked potential (SSVEP) analysis, which separates out the responses of photoreceptors, lamina neurons, and medulla neurons, and highlights any changes in synaptic transmission in the retina (Afsari et al., 2014). Flies were raised in the dark or in a “disco-chamber” where the light is turned on and off at random to induce a mild visual stress (Fig. 6 d). This stress accelerates neurodegeneration (Hindle et al., 2013), thereby mimicking neuronal loss in PD. Flies expressing LRRK2 G2019S exhibited significantly smaller visual responses in the lamina neurons and photoreceptors after disco (Fig. 6 d). In contrast, visual responses were not perturbed in flies raised in the dark (Fig. 6 d). Visual responses in two independent TPC2-expressing lines were also substantially reduced in the light but not in the dark, thereby phenocopying LRRK2 G2019S (Fig. 6 d). We additionally created and analyzed double transgenics expressing both TPC2 and LRRK2 G2019S. Importantly, visual disturbances were readily detected in the double-transgenic flies reared in the dark (Fig. 6 d). Thus, whereas TPC2 and LRRK2 G2019S can mediate dopaminergic dysfunction independently of each other under stressed conditions, they synergize in a more physiologically unstressed setting. TPC2 thus recapitulates the deleterious actions of pathogenic LRRK2 on dopaminergic function in vivo.

Movement was assayed in two ways. In the first, we leveraged the proboscis extension reflex to sucrose, a simple circuit driven by motoneurons that results in the contraction of the proboscis muscle (Fig. 6 e). The circuit is modulated by a single dopaminergic neuron (TH-VUM) that is spontaneously active (much like dopaminergic neurons in the mammalian substantia nigra) and that synapses with the sensory neurons and interneurons (Afsari et al., 2014; Cording et al., 2017; Fellgett et al., 2021). The reflex was assessed by quantifying the proportion of starved flies that extended their proboscis in response to a sugar stimulus. As summarized in Fig. 6 f, nearly all control flies, either young (1 day) or middle aged (4 days), responded to sucrose. This proportion was reduced in aged flies (7 days) to approximately two-thirds. In flies expressing LRRK2 G2019S, locomotor activity was reduced in both middle-aged and aged flies relative to control flies (Fig. 6 f). Essentially, similar results were obtained in transgenic flies expressing TPC2 alone or in combination with G2019S LRRK2 (Fig. 6 f). The proportion of responsive flies was reduced in two independent G2019S-expressing lines and three independent TPC2-expressing lines (Fig. S5, a and b). This defect was particularly prominent in middle-aged double transgenics (Fig. 6 f) pointing to an interaction between LRRK2 and TPC2 as evidenced in our visual assays. Thus, TPC2 again phenocopies and likely synergizes with LRRK2 G2019S to perturb movement in vivo.

The targeted expression of tetanus toxin (Suster et al., 2003) or pathogenic alpha-synuclein (Varga et al., 2014) in dopaminergic neurons inhibits locomotor behavior of larvae. Therefore, in an independent approach to assess movement, we examined the effect of TPC2 and/or LRRK2 G2019S on larval crawling (Fig. 6 g). As shown in the representative tracks in Fig. 6 h, larvae could be readily monitored and locomotion quantified by measuring the distance traveled over 60 s. In larvae expressing LRRK2 G2019S, crawling was reduced relative to control larvae (Fig. 6 h). Similar results were obtained with the R1441C mutation in LRRK2 (Fig. S5, c–e). These effects were specific because the expression of wild-type LRRK2 or kinase-inactive G2019S LRRK2 did not affect crawling (Fig. S5, c–e).

Crawling was also reduced in larvae expressing TPC2 (Fig. 6 h). As with the vision responses to disco, the deleterious effects of LRRK2 G2019S and TPC2 were not additive based on movement in the double-transgenic flies. These data quantified in Fig. 6 i provide further in vivo evidence that TPC2 mimics LRRK2 G2019S. The expression of TPC2 was readily detectable in the larvae by confocal microscopy where it colocalized with LAMP1 (Fig. 6 j). We measured cytosolic Ca²⁺ in brain explants using jGCaMP8m (Fig. S5 f) and resolved spontaneous activity in dopaminergic neurons from control larvae and larvae expressing LRRK2 G2019S and TPC2 (Video 1 and Fig. S5 g). These signals are consistent with autonomous activity in the larval brain.

Taken together, we show that TPC2, like LRRK2 G2019S, is sufficient to disrupt dopaminergic circuits underpinning both vision and movement in vivo.

To explore the underlying mechanisms underpinning TPC2-dependent dopaminergic neuron dysfunction in vivo, we generated flies expressing a series of TPC2 mutants (Fig. 7 a) that included the pore-dead and plasma membrane–targeted variants.

The corresponding visual responses are shown in Fig. 7 b. These analyses revealed that whereas wild-type TPC2 perturbed vision, the pore-dead mutant that is demonstrably inactive in vitro (Fig. 5) did not. These data indicate that channel activity is required for TPC2 action in vivo. To mitigate against nonspecific effects of channel activity, we took two approaches. In the first, we leveraged TPC2 rerouted to the plasma membrane, which is demonstrably active but mistargeted (Fig. 5). In the second, we examined the effects of TPC1, which is closely related to TPC2 but expressed in earlier compartments of the endo-lysosomal system (Brailoiu et al., 2009b). Both TPC2^PM and TPC1 failed to disrupt vision in vivo (Fig. 7 b). These data suggest that dopaminergic dysfunction results from aberrant channel activity specifically on the lysosome.

Much evidence implicates Rab proteins in LRRK2 action, and TPC2 is a Rab interactor (Lin-Moshier et al., 2014). We therefore analyzed TPC2 in which the Rab binding site within the N terminus of TPC2 was mutated (Fig. 7 a). This mutant (TPC2^RD, for Rab-dead) failed to perturb dopaminergic function (Fig. 7 b), suggesting that the inhibitory effects of TPC2 require Rab interactivity. Activation of Na⁺ currents through TPC2 by PI(3,5)P₂ proceeds through residues in the S4-S5 linker. We therefore also analyzed the effect of mutating Lys 204 (Fig. 7 a), which is PI(3,5)P₂-insensitive (She et al., 2019). TPC2^LD (for lipid-dead) disrupted vision similar to wild-type TPC2 (Fig. 7 b), suggesting that the effects of TPC2 are PI(3,5)P₂-independent and thus likely driven by Ca²⁺.

In an independent approach, we examined the effects of the TPC2 mutants on larval crawling (Fig. 7 c). Similar to the vision assays, the TPC2 pore and targeting mutants were without effect on movement (Fig. 7 d). So too were TPC1 and the TPC2 Rab-dead mutant (Fig. 7 d). In further analogy with the effects on vision, the lipid-dead TPC2 mutant produced inhibitory effects (Fig. 7 c). Thus, there was marked congruence in the neurophysiological and behavioral phenotypes whereby TPC2 but not TPC1 Ca²⁺ channel activity in the lysosome disrupts vision and movement in a Rab-dependent manner. Confocal microscopy of larval brains confirmed the expression of TPC2 in all lines (Fig. S5 h).

Finally, we examined whether chemically targeting TPC2 could improve dopaminergic function in vivo. For these experiments, larvae expressing TPC2 were fed with TPC2-A1-P, which reverses Ca²⁺ defects in vitro (Fig. 4), and its effect on crawling was assessed. As shown in Fig. 7 e, including TPC2-A1-P in food through development reversed the inhibitory effects of TPC2 on movement at the larval stage. In contrast, feeding with TPC2-A1-N did not (Fig. 7 e). The effect of TPC2-A1-P on crawling was dose-dependent (Fig. 7 f). The normal absence of TPC2 in Drosophila allowed us to directly probe off-target effects of TPC-A1-P. As shown in Fig. 7 f, TPC2-A1-P had little effect in control larvae firmly attesting to specificity. Thus, dopaminergic dysfunction induced by TPC2 in vivo can be reversed by pharmacologically targeting TPC2 in an agonist-selective manner.

Overall, we provide mechanistic insight into and a chemical reversal strategy for TPC2-dependent dopaminergic dysfunction in vivo.

Here, we identified Ca²⁺, neurophysiological, and behavioral defects mediated by TPC2 and showed that targeting this channel reversed phenotypes mediated by a PD-relevant mutation in LRRK2.

Our data revealed that voltage-gated Ca²⁺ entry is increased upon mutation of LRRK2 in a model dopaminergic neuron (Fig. 1), as well as iPSC-derived midbrain dopaminergic neurons (Fig. 4). Changes in voltage-gated Ca²⁺ channel expression in PD have been known for some time (Hurley and Dexter, 2012). More recent studies have identified increased Ca²⁺ levels and increased voltage-gated Ca²⁺ currents in LRRK2 G2019S–expressing dopaminergic neurons derived from iPSCs and knock-in mice resulting from increased translation (Kim et al., 2020, 2021). Enhanced evoked Ca²⁺ entry in the presence of the LRRK2 G2019S mutation reported here supports these findings. Indeed, growing literature identifying defects in iPSC-derived dopaminergic neurons is consistent with a cell-autonomous role for LRRK2 in regulating neuronal activity (Bono et al., 2021; Carola et al., 2021; Kim et al., 2020). Our data, however, contrast results in iPSC-derived sensory neurons where K⁺-induced Ca²⁺ entry is reduced by the G2019S mutation (Schwab and Ebert, 2015) and in a mixed iPSC-derived neuronal population where K⁺-induced Ca²⁺ entry is enhanced but only after depletion of endoplasmic reticulum Ca²⁺ stores (Korecka et al., 2019). This suggests the LRRK2-induced defects may manifest in a cell type–specific manner. Enhanced Ca²⁺ entry was unlikely to be due to compromised endoplasmic reticulum Ca²⁺ homeostasis or cytoplasmic buffering because neither the Ca²⁺ signals evoked upon endoplasmic reticulum Ca²⁺ depletion nor those upon subsequent Ca²⁺ entry were affected upon LRRK2 mutation. Our data, using a number of chemical and molecular interventions, including the use of two chemically distinct permeabilizing agents (Figs. 2 and S2) (Atakpa et al., 2019; Morgan et al., 2020; Yuan et al., 2021), instead identified a prominent role for lysosomes and TPC2 in mediating Ca²⁺ defects at the plasma membrane.

In hippocampal neurons, voltage-gated Ca²⁺ influx is functionally coupled with NAADP-mediated lysosomal Ca²⁺ release (Padamsey et al., 2017). Thus, as reported here, antagonizing NAADP action reduced postdepolarization-induced Ca²⁺ influx. Interestingly, this interplay was coupled with lysosomal exocytosis to maintain dendritic morphology (Padamsey et al., 2017)—a key process regulated by LRRK2 (MacLeod et al., 2006) and NAADP (Brailoiu et al., 2005). Further evidence that lysosomal Ca²⁺ release and external Ca²⁺ entry are coupled comes from our data using recently described TPC2 agonists to directly activate TPC2 (Gerndt et al., 2020). We show that the NAADP-mimetic TPC2-A1-N evokes Ca²⁺ signals with a major Ca²⁺ entry component when combined with a PI(3,5)P₂ mimetic (which alone does little to Ca²⁺) (Figs. 3 and 4). Such synergism adds to similar effects of TPC2 activation observed in cancer cells and pancreatic acinar cells serving as a key functional diagnostic of endogenous TPC2 channels (Yuan et al., 2022). Local NAADP-evoked Ca²⁺ signals have been reported to depolarize medulla neurons (Brailoiu et al., 2009b). Thus, lysosome–plasma membrane crosstalk is likely bidirectional with Ca²⁺ entry evoking lysosomal Ca²⁺ release and lysosomal Ca²⁺ release evoking Ca²⁺ entry. More work is required to understand exactly how TPC2 and voltage-gated Ca²⁺ channels communicate within this novel axis. Efforts at targeting Cav1.3 are underway in PD (Caulfield et al., 2023). We suggest that targeting TPC2, perhaps in combination, also be considered.

TPCs are ancient ion channels that can be traced back to unicellular organisms (Li et al., 2021; Rahman et al., 2014). But there has been lineage-specific loss of TPC isoforms in mammals (Brailoiu et al., 2010a; Cai and Patel, 2010) and, as reported here, complete loss in some insects (Fig. 5). The absence of TPCs in Drosophila appears to be a result of chromosomal fragmentation and an exception rather than the rule in insects since TPCs were identifiable in other insect orders. We leveraged this loss to study TPC2 heterologously expressed in a “clean” background and importantly its pharmacology in a genetically tractable way. But it should be noted that we cannot rule out functional compensation of TPC loss by other lysosomal channels. Our data show that the expression of TPC2 in dopaminergic neurons disrupted vision and movement, thereby phenocopying pathogenic LRRK2. Our in vivo electrophysiological assays centered around visual stress evoked by repetitive light episodes (“disco”). But the dark controls proved informative too as TPC2 and LRRK2 synergized to disrupt dopaminergic function, thereby placing the two proteins in a common pathway in the eye. Notably, we established crawling assays in flies expressing TPC2 and LRRK2 G2019S (or both) and found locomotion defects (Fig. 6). At present, we know little about the function of TPCs in a neuronal context but our studies add to recent ones implicating them in synaptic plasticity (Foster et al., 2018) and social behavior (Martucci et al., 2023) building on early work identifying a role for NAADP in neurotransmission (Brailoiu et al., 2003) and neuronal differentiation (Brailoiu et al., 2006).

To explore the mechanisms underpinning TPC2-mediated defects, we translated point mutations in TPC2 with defined action at the channel level through to dopaminergic deficits in vivo (Fig. 7). Our data revealed five salient features: first, TPC2 disruptions were pore-dependent consistent with reversal of LRRK2 Ca²⁺ defects by the same mutant (Fig. 2). Second, the effects required localization to the lysosome, indicating the highly localized fluxes as the driver as opposed to nonspecific channel activity. Third, defects were specific for TPC2 over TPC1 further attesting to specificity and further suggesting localized activity given that TPC1 is targeted to more distal compartments of the endo-lysosomal system than TPC2 (Brailoiu et al., 2009a). Fourth, disruption may require Rab interactivity. The Rab family of trafficking proteins are of particular relevance to LRRK2 signaling (Kuwahara and Iwatsubo, 2020). Similar to TPC2, LRRK2 reportedly interacts with a number of these molecular switches including Rab5 (Shin et al., 2008), Rab7 (Dodson et al., 2012), and Rab7L1/Rab29 (MacLeod et al., 2013). The identification of Rab8 and Rab10 as direct LRRK2 substrates (Steger et al., 2016) further implicates LRRK2 in regulating endo-lysosomal morphology and function (Eguchi et al., 2018). Fifth, disruption was independent of PI(3,5P)₂ binding. PI(3,5)P₂-mediated Na⁺ signaling is abolished by the K204A mutation, but this did not prevent functional deficits. Thus, Na⁺ flux is probably not relevant here pointing instead to Ca²⁺ as the culprit. Recent work has shown that the mutation of K204 and other residues that form the PI(3,5P)₂ binding site also block NAADP activation (Saito et al., 2023). NAADP normally induces Ca²⁺ flux through TPC2 via NAADP binding proteins (Gunaratne et al., 2021, 2023; Marchant et al., 2022). Thus, TPC2^LD might be considered “ligand-dead” as opposed to lipid-dead, meaning that it is constitutive leak of Ca²⁺, not evoked flux through TPC2, which is (dys)functionally relevant. We found that various mutants had essentially indistinguishable results in our vision and movement assays. We therefore propose this approach as a novel and robust systems-level opportunity for interrogating TPC2 functionality.

Ostensibly, the most interesting finding reported here is reversal of defects by targeting of TPC2. For LRRK2-Ca²⁺ defects, antagonizing NAADP action and blocking the TPC2 pore either chemically or molecularly were sufficient to reset Ca²⁺ signals (Fig. 2). These data extend previous work in fibroblasts from LRRK2 G2019S patients where lysosomal morphology defects were reversed by targeting of TPC2 but not TPC1 (Hockey et al., 2015) to a neuronal setting. But equally, activating TPC2 with TPC2-A1-P also reset deviant Ca²⁺ signals, both in neuroblastoma cells overexpressing LRRK2 G2019S and in iPSC-derived dopaminergic neurons expressing pathogenic LRRK2 at endogenous levels (Fig. 4), as well as movement defects in vivo (Fig. 7). This was despite TPC2-A1-P synergizing with TPC2-A1-N to mediate Ca²⁺ influx (Figs. 3 and 4). TPC2-A1-P was shown recently to reverse lysosomal defects in iPSC models of several lysosomal storage disorders and an in vivo mouse model of Mucolipidosis IV (Scotto Rosato et al., 2022). Given the link between lysosomal storage diseases and PD, we speculate that lysosomal Ca²⁺ is a common defect, or, in the very least, a common target to improve well-being in these disorders. Lysosomal Ca²⁺ defects were reported in GBA1-linked PD (Kilpatrick et al., 2016) building on foundational studies in models of Niemann–Pick type C disease (Lloyd-Evans et al., 2008). Kufor–Rakeb syndrome is a PD-like syndrome caused by mutation of a ATP13A2 (Ramirez et al., 2006), which encodes a lysosomal ATP-dependent pump with a possible role in Ca²⁺ transport (Narayanaswamy et al., 2019). Targeting TPC2 as described here may thus be of broader benefit in tackling PD. Beyond PD, lysosomal Ca²⁺ defects have been noted in models of Alzheimer’s disease and inhibiting TPC2 provides benefit (Tong et al., 2022). We suggest that selective biasing of TPC2 away from deleterious Ca²⁺ fluxes might have less side-effects than pan-inhibition and propose that such a strategy might find utility in tackling neurodegeneration.

SH-SY5Y cells were maintained in 1:1 mixture of DMEM and Ham’s F12 medium, supplemented with 10% vol/vol fetal bovine serum (FBS), 100 U/ml penicillin, 100 µg/ml streptomycin, and 1% nonessential amino acids (all from Gibco) at 37°C in a humidified 5% CO₂ atmosphere. Stable cell lines expressing wild-type LRRK2, pathogenic LRRK2 G2019S, and kinase-inactive LRRK2-K1906A/D1994A/D2017A were generated as described previously (Papkovskaia et al., 2012). Differentiation was induced 24 h after plating by the addition of 10 µM retinoic acid (from Sigma-Aldrich) to 1:1 mixture of DMEM and Ham’s F12 medium supplemented with 2% vol/vol FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin (all from Gibco). Cells were differentiated for 5–7 days.

iPSC lines were obtained from Parkinson’s Progression Markers Initiative. iPSCs were generated from peripheral blood mononuclear cells obtained from healthy controls and people with PD carrying the LRRK2 G2019S mutation (Table S1). Cells were cultured in Essential 8 media (Thermo Fisher) on truncated human vitronectin (Thermo Fisher) in 6-well plates.

S2R⁺ cells (obtained from the Drosophila Genomics Resource Centre, Indiana University, Bloomington, IN, USA) were maintained in Schneider’s Drosophila medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin (all from Gibco) at 25°C as semi-adherent cultures.

Cells were plated onto either a 6-well plate, a T25 flask, or 13-mm glass coverslips coated with 20 µg/ml poly-L-lysine (Sigma-Aldrich) in a 24-well plate for qPCR analysis and imaging. For automated Ca²⁺ measurements, cells were plated onto opaque-walled 96-well microplates (Corning).

RNA was extracted using RNeasy Mini Kit and RNase-free DNase Set (both from Qiagen). cDNA was synthesized using Superscript III Reverse Transcriptase (Invitrogen), Random Primers (Promega), and Oligo dT(12-18) primers (Invitrogen). PCR was performed using SYBR Green JumpStart Taq Ready Mix (Sigma-Aldrich) and primers for human LRRK2 and the housekeeping gene, UBC. Samples were denatured for 2 min at 94°C, followed by 40 cycles of denaturation (15 s at 94°C), annealing (30 s at 60°C), and extension (30 s at 72°C). LRRK2 expression was normalized to UBC expression. Details of the primers used are listed in Table S2. Western blotting (Papkovskaia et al., 2012) was performed using antibodies to LRRK2 (1:1,000, ab133518; Abcam) and cathepsin D (1:1,000, ab6313; Abcam) (Papkovskaia et al., 2012).

Mammalian expression plasmids encoding GECO1.2 (Zhao et al., 2011), TPC2^L265P-GFP (Brailoiu et al., 2010b), and LAMP1-GFP (Falcón-Pérez et al., 2005) were described previously.

Insect expression plasmids encoding TPC2 were based on pUASTattB (Bischof et al., 2007). For TPC2-GCaMP, the coding sequence of human TPC2 fused to GCaMP6s (Gerndt et al., 2020) was amplified by PCR using the primers described in Table S3, and the product was inserted into the EcoRI and XbaI sites of pUASTattB. TPC2^L265P-GCaMP was generated by site-directed mutagenesis of pUASTattB TPC2-GCaMP using the primers described in Table S3. For TPC2-mCherry, the coding of sequence of TPC2 GFP in pCS2+ (Brailoiu et al., 2009a) was directly inserted into the EcoRI and XbaI sites of pUASTattB and the GFP coding sequence at the XhoI and XbaI sites replaced with mCherry. The coding sequences of TPC2^L265P (Brailoiu et al., 2010b), TPC2^L11A/L12A (Brailoiu et al., 2010b), and TPC2^{Q33A/V34A/P36A} (Lin-Moshier et al., 2014) were directly inserted into the EcoRI and XhoI sites of pUASTattB TPC2-mCherry to form the corresponding mCherry-tagged constructs. TPC2^K204A-mCherry was generated from TPC2-mCherry by site-directed mutagenesis using the primers described in Saito et al. (2023). For TPC1-mCherry, the coding sequence of TPC1 (Brailoiu et al., 2009a) was amplified by PCR using the primers listed in Table S3, and the product was inserted into the EcoRI and NotI sites of pUASTattB. mCherry was inserted at the NotI and XbaI sites. Drosophila LAMP fused to GFP was described in Minin et al. (2006).

SH-SY5Y cells were transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s protocol and used 12–24 h after transfection. S2R⁺ cells were transfected using the Effectene transfection reagent kit (Qiagen) according to the manufacturer’s instructions and used 24–48 h after transfection. GAL4-dependent UAS expression plasmids were cotransfected at a 1:1 ratio with a pWAGal4 (Kuranaga et al., 2006) (a kind gift from Dr. Nic Tapon, Francis Crick Institute, London, UK).

iPSCs were differentiated into midbrain dopaminergic neurons using the floor-plate protocol described in Kriks et al. (2011).

Briefly, iPSCs were passaged with Accutase to yield a single-cell suspension and seeded at 1 × 10⁶ cells/ml in STEMdiff SMADi Neural Induction Media (Stemcell Technologies) supplemented with 10 mM Y27632 (Abcam) on Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix (Thermo Fisher). Media were changed daily for 15 days with one passage at day 6 or 7. Midbrain floor-plate neural precursor cells were then generated by passaging neuroepithelial cells with Accutase and resuspending in 1:1 mixture of N2 media (DMEM:F12 media containing 1X N2 supplement, 1X nonessential amino acids, 10 mM β-mercaptoethanol, 5 mg/ml insulin) and B27 media (Neurobasal media containing 1X B27 supplement, 1X GlutaMAX). The N2:B27 medium mix was supplemented with 20 ng/ml BDNF, 100 ng/ml FGF8, 1 ng/ml TGFβ3 (all from PeproTech), 0.1 mM compound E (Abcam), 200 mM ascorbic acid, and 0.5 mM dibutyryl cAMP (Enzo Lifesciences). Cells were seeded at 0.75 × 10⁶ cells/ml on Geltrex-coated plates and media changed daily for 4 days; and then every other day (no FGF8) for a further 4 days. Neural precursor cells were then passaged with Accutase and seeded for terminal differentiation at 0.125 × 10⁶ cells/ml on polyornithine and laminin (1 mg/ml)-coated coverslips in BrainPhys neuronal media containing SM1 (Stemcell Technologies), and 1X N2 supplement, and supplemented fresh with 20 ng/ml BDNF, 20 ng/ml GDNF, 1 ng/ml TGFβ3, 0.1 mM compound E, 40 mM ascorbic acid, and 0.5 mM dibutyryl cAMP. Cells were cultured for 8 wk with half-medium changes every 2 or 3 days.

The genomic locations of insect TPC genes were identified using TBlastN searches (Altschul et al., 1997) with the TPC protein sequence from Ctenocephalides felis (XP_026472147). The initial screening of syntenic chromosomal regions surrounding TPC homologs was made among the genomes of Ctenocephalides felis, Bombyx mandarina, Ceratina calcarata, and Pieris brassicae. The genes of 60S ribosomal protein L6 (RPL6), mevalonate kinase (MVK), bolA-like protein 2 (BOLA2), and dolichyl pyrophosphate Man9GlcNAc2 alpha-1,3-glucosyltransferase (ALG6) were used for further analysis of synteny comparison flanking TPC homologs among select insect genomes.

SH-SY5Y cells and iPSC-derived midbrain dopaminergic neurons plated on coverslips were loaded with the ratiometric fluorescent Ca²⁺ indicator Fura-2 by incubation for 1 h at room temperature (RT) with the acetoxymethyl ester form of Fura-2 (Fura-2 AM, 2.5 µM) and 0.005% (vol/vol) pluronic acid (both from Invitrogen) in HEPES-buffered saline (HBS) containing 10 mM HEPES, 1.25 mM KH₂PO4, 2 mM MgSO₄, 3 mM KCl, 156 mM NaCl, 2 mM CaCl₂, and 10 mM glucose (all from Sigma-Aldrich) at pH 7.4. S2R+ cells were loaded similarly except for the addition of 2.5 mM probenecid to prevent dye leakage. After labeling, cells were quickly washed three times in HBS and mounted in an imaging chamber (Bioscience Tools) with 1 ml HBS.

Epifluorescence images were captured every 3 s with a cooled coupled device camera (IMAGO; TILL Photonics) attached to an Olympus IX71 inverted fluorescence microscope that was fitted with a UApo/340 20×/0.70 W ∞/0.17 objective and a monochromatic light source (Polychrome IV) under the control of TILLvisION v4 software. The Fura-2 ratio was recorded using dual-excitation wavelengths of 340 and 380 nm, and a 440-nm long-pass filter was used to collect emitted fluorescence. Cells expressing GFP-tagged constructs were identified by excitation at 488 nm and capturing emitted fluorescence using a 515-nm long-pass filter. Cells expressing mRFP-tagged constructs were identified by excitation at 568 nm and capturing emitted fluorescence using a 590-nm long-pass filter. For some experiments, epifluorescence images were captured using a Megapixel monochrome cooled coupled device camera attached to an Olympus IX73 inverted fluorescence microscope that was fitted with a UPlanXApo 20×/0.80 ∞/0.17/OFN26.5 objective and a CoolLED multiple wavelength LED source under the control of MetaFluor 7.10.3.279 software. Fura-2 was excited at 340/380 nm, and emitted fluorescence was captured using a 425-nm long-pass filter.

Automated measurement of cytosolic Ca²⁺ in SH-SY5Y cells in a 96-well plate was performed using a fluorescence plate reader (CLARIOstar, BMG Labtech) under the control of Mars 3.42 R3 software. A single measurement comprised 16 flashes at 335 and 380 nm (each at 8-nm band-pass [BP]) while recording fluorescence at 520 nm (90-nm BP). Measurements were repeated on an individual well at 3-s intervals with 15 wells being recorded in parallel using “plate mode.” Defined volumes of HBS and modified HBS containing 120 mM KCl in place of NaCl were injected simultaneously through two independent injector needles to achieve final concentrations in the range of 10–90 mM KCl. Background fluorescence was measured from wells containing cells that were incubated with HBS without Fura-2.

SH-SY5Y cells plated on coverslips that transiently expressed GECO1.2 were mounted in an imaging chamber (Bioscience Tools) with 1 ml HBS and epifluorescence images captured every 3 s with a cooled coupled device camera (IMAGO; TILL Photonics) attached to an Olympus IX71 inverted fluorescence microscope that was fitted with a UApo/340 40×/1.35 Oil Iris ∞/0.17 objective and a monochromatic light source (Polychrome IV) under the control of TILLvisION v4 software. GECO1.2 was excited at 470 nm, and emitted fluorescence was collected using a 515-nm long-pass filter.

S2R+ cells plated on coverslips that transiently expressed GCaMP6s-tagged TPC2 constructs were mounted similarly and epifluorescence images captured every 3 s with a Megapixel monochrome cooled coupled device camera attached to an Olympus IX73 inverted fluorescence microscope that was fitted with a UPlanXApo 20×/0.80 ∞/0.17/OFN26.5 objective and a Cairn CoolLED multiple wavelength LED source under the control of MetaFluor 7.10.3.279 software. GCaMP6s was excited at 470 nm, and emitted fluorescence was collected using a 510-nm long-pass filter.

In some experiments, cells were treated overnight with either 0.1% (vol/vol) DMSO (Sigma-Aldrich) or 1 µM vacuolin-1 (Chem Cruz) before being loaded with Fura-2. In other experiments, 1 mM LLOMe (Cayman Chemical Company) was included during loading. Either 0.1/0.2% (vol/vol) DMSO (Sigma-Aldrich) or H₂O was used as controls when appropriate. SH-SY5Y cells were depolarized by replacing HBS with high-potassium buffer, containing 10 mM HEPES, 1.25 mM KH₂PO4, 2 mM MgSO4, 156 mM KCl, 3 mM NaCl, 2 mM CaCl₂, and 10 mM glucose (all from Sigma-Aldrich) at pH 7.4 such that the final KCl concentration was 50 mM. Cells were also acutely challenged with ionomycin, thapsigargin, PF-543, tetrandrine, ML-SI3 (synthesized as described previously [Leser et al., 2021]), riluzole and the TPC2 agonists, TPC2-A1-N, TPC2-A1-P, and SGA-10 (all synthesized as described previously [Gerndt et al., 2020]). Some stimulations were performed in nominally Ca²⁺-free HBS where Ca²⁺ was omitted from HBS.

To monitor lysosome integrity and morphology, SH-SY5Y cells plated on coverslips were loaded with 0.1 mg/ml dextran-conjugated rhodamine B (MW 10,000; Invitrogen) overnight at 37°C in culture. Cells were subsequently washed in dextran-free complete medium and cultured for a further 4 h (chase period) to label lysosomes. Epifluorescence and transmitted light images were captured with a cooled coupled device camera (IMAGO; TILL Photonics) attached to an Olympus IX71 inverted fluorescence microscope that was fitted with a UApo/340 20×/0.70 W ∞/0.17 objective and a monochromatic light source (Polychrome IV) under the control of TILLvisION v4 software. Rhodamine B was excited at 568 nm, and emitted fluorescence was captured with a 590-nm filter.

Transfected SH-SY5Y and S2R⁺ cells plated on coverslips were fixed for 15 min with 4% (vol/vol) paraformaldehyde (VWR) prepared in phosphate-buffered saline (PBS; Sigma-Aldrich) at RT. After three 5-min washes in PBS, nuclei were labeled by incubation for 5 min with 1 µg/ml 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) at RT. Coverslips were washed thrice for 5 min with PBS, mounted onto microscope slides with 1,4-diazabicyclo[2,2,2]octane (Sigma-Aldrich), and sealed with colorless transparent nail varnish.

In some experiments, SH-SY5Y cells were stained for endogenous LAMP1 (Saito et al., 2023). Cells were fixed with paraformaldehyde, permeabilized with β-escin, blocked with bovine serum albumin/FBS, and sequentially incubated with antibodies and DAPI with intervening washes with Tween-20 prior to mounting using Fluoromount-G. Antibodies used were mouse LAMP1 antibody (H4A3) (1:10 dilution; catalog no. sc-20011 from Santa Cruz Biotechnology) and donkey anti-mouse (Alexa Fluor 594; 1:100 dilution; catalog no. A21203 from Thermo Fisher Scientific). In other experiments, S2R+ cells were incubated for 2 h at 37°C with 0.01% BioTracker 490 Green Cytoplasmic Membrane Dye (Merck) after fixation.

Confocal images of SH-SY5Y cells were taken using an inverted Axio Observer.Z1 microscope attached to a VivaTome scanner (Zeiss) fitted with an LCI Plan-Neofluar 63×/1.3 Imm Corr DIC M27 objective. GECO and GFP were excited using a 432- to 482-nm BP filter, and emitted fluorescence was captured using a 500- to 550-nm BP filter.

Confocal images of SH-SY5Y cells subject to immunocytochemistry, and S2R+ cells were acquired using an inverted Axio Observer.Z1 microscope attached to a LSM 880 confocal scanner (Zeiss) fitted with a Plan-Apochromat 63×/1.4 Oil DIC objective. DAPI, GFP/BioTracker 490 Green, and mCherry fluorescence were excited using wavelengths of 405, 488, and 561 nm, respectively. Emitted fluorescence was captured using either 410–479-nm, 490–577-nm, or 578–696-nm BP filters, respectively. 16-bit images were taken at 4× optical zoom.

All confocal images were acquired using ZEN Black software and processed on Fiji (ImageJ) software. For colocalization analyses, negative controls comprised the same image, but with the red channel rotated 180° to the right. Pearson’s correlation coefficient was obtained from confocal images using the “Just Another Co-localization Plugin” (JACoP).

Drosophila were raised on standard cornmeal–yeast–sucrose medium at 25°C on a 12-h light:dark cycle.

UAS-TPC2-mCherry was created in pUAST-AttB (Bischof et al., 2007) and injected into the AttP2 site (RRID:BDSC_25710) at the NCBS Drosophila Facility (Bangalore). All other TPC2 constructs were injected by the Drosophila Injection Service at FlyORF (Zurich). TH-GAL4 (Friggi-Grelin et al., 2003) was a kind gift from Serge Birman (ESPCI, Paris, France), UAS-Lrrk2 and UAS-Lrrk2-G2019S (Liu et al., 2008) were kind gifts from Wanli Smith (Johns Hopkins, Baltimore, MD, USA), UAS-Lrrk2 R1114C and UAS-Lrrk2-G2019S/K1906M (Lin et al., 2010) were kind gifts from Cheng-Ting Chien (Academia Sinica, Taipei, Taiwan), UAS-Lamp1-GFP (Pulipparacharuvil et al., 2005) was a kind gift from Helmut Kramer (UT Southwestern, Dallas, TX, USA), and UAS-jGCaMP8m (Zhang et al., 2023) was a kind gift from James Jepson (UCL, London, UK). As a control, TH-GAL4 flies were crossed to the AttP2 landing site stock (empty vector), Bloomington Stocks Centre (RRID:BDSC_36303).

On the day of emergence, flies were placed in the dark or in disco-chambers at 29°C. 1-wk-old flies were prepared for SSVEP using a pooter and nail polish to secure them in a cutoff pipette tip, without anesthesia. Each fly was presented five times with a set of nine flickering stimuli. In each stimulus, the average light intensity was the same, but the amplitude of the flicker was adjusted from 10% to 100%, giving a range of contrasts. Fast Fourier transform (FFT) was applied to the responses, to separate the first harmonic (1F1), due to the photoreceptors, from the second harmonic (2F1), due to the lamina neurons. Stimuli were generated and responses recorded by an Arduino Due system with FFTs and contrast sensitivity computed in MATLAB. https://github.com/wadelab/flyCode

Behavior was recorded from 1- to 7-day-old flies kept in the dark at 29°C. Flies were collected under CO₂ anesthesia, and restrained by sticking them ventral side up to card with rubber cement (Fixo Gum). Flies were left to recover for 3 h at 29°C. They were presented with a droplet of 100 mM sucrose solution to the legs, and the immediate response (<2 s) was scored.

Third instar wandering larvae were immersed in PBS, on a Sylgard (DuPont) plate pinned at the anterior and posterior and the dorsal body wall cut along the midline from posterior to anterior. Fat bodies and internal organs were removed while leaving the CNS in place. PBS was then removed and replaced with 3.7% formaldehyde/PBS for 7 min. Preparations were then washed in PBT three times and then immersed in 70% glycerol/PBS for 2 h. Preparations were then mounted in Vectashield (Vectorlabs) and images collected from expressing dopaminergic neurons in the larval CNS on a Zeiss LSM 880 confocal microscope (inverted) with 10×/0.3 air or 63×/1.4 oil objectives. GFP was excited using 488-nm laser line and emission collected at 490–550 nm. mCherry was excited using 561-nm laser line and emission collected at 570–650 nm.

Larval crawling assays were performed at 25°C using third instar wandering larvae. Two to three larvae were transferred onto the center of a 90-mm-diameter petri dish containing a thin layer of 1% agar and left to acclimatize for 1–2 min. The petri dish was placed upon a black surface and imaged from above using an iPhone 11 recording at 30 frames per second (fps) for 60 s. Videos were converted from .MOV to .MP4 format using HandBrake (https://handbrake.fr), with dimensions converted from 1080 × 1920 to 720 × 1280 pixels and 5 fps. Recordings were then converted to .avi format using ffmpeg (https://ffmpeg.org). Videos were then imported and analyzed in ImageJ using the MTrack2 plugin. Images were converted to 8-bit images and grayscale and the background subtracted to observe the larvae as white objects on a black background. Total track length was measured. For experiments testing the effects of TPC2 agonists, Formula 4-24 Instant Medium (Carolina Biological Supply) was supplemented with the drugs added from ethanol stocks.

Intact third instar central nervous system (mouth hooks, eye-antennal disks, optic lobes, central brain, and ventral nerve cord) was quickly dissected in ice-cold sterile explant culture medium (∼4°C), consisting of Schneider’s Drosophila medium (Sigma-Aldrich) supplemented with 2.5 µg/ml human insulin (Sigma-Aldrich), 1% penicillin–streptomycin (Sigma-Aldrich), and 10% FBS (Sigma-Aldrich), under a dissecting microscope (Zeiss Stemi 2000). Explants were immobilized in an agarose scaffold as described previously (Bostock et al., 2020). Briefly, 1 ml 0.4% wt/vol low-temperature gelling agarose (Sigma-Aldrich) mixed with explant culture medium was heated to 42°C. The temperature of the agarose medium was then lowered to 34°C and pipetted into petri dishes (35 × 10 mm; Thermo Fisher Scientific) to which a single brain explant was added and oriented. The agarose was left to solidify for 15 min and the dish filled with sterile ice-cold explant culture medium comprising Schneider’s Insect medium (#S0146; Sigma-Aldrich) supplemented with 2.5 μl/ml human insulin (#I9278; Sigma-Aldrich), 1 % penicillin–streptomycin (#P4333; Sigma-Aldrich), and 10 % FBS (#F2442; Sigma-Aldrich).

Time-lapse Z-stacks were acquired using an upright Zeiss LSM 880 microscope with Spectra-Physics Mai Tai DeepSee two-photon lasers fitted with a W Plan-Apochromat 20×/1.0 DIC M27 70-mm objective. jGCaMP8m was excited at 935 nm (laser power <15%) and emission collected between 500 and 550 nm. 8-bit Z-stacks (10-40 slices per stack) were obtained with a Z interval of 1 µm and a scan time of 6–26 s per slice using ZEN Black software. Imaging was performed at 25°C.

Icy and ImageJ software were used to analyze movies. ROIs were manually drawn within every jGCaMP8m-expressing cell. The ROI intensity evolution plugin was used to extract the mean intensity evolution over time.

Data are presented as representative results from a single technical replicate (i.e., coverslip) or as the mean ± SEM from three or more independent experiments. Statistical analysis was performed using Prism 9 on datasets consisting of independent experiments performed on three or more different days per group/condition (i.e., n ≥ 3).

For two-group comparisons, an unpaired t test was used. For multiple groups, a one-way or two-way ANOVA followed by either Tukey’s, Sidak’s, or Dunnett’s tests was used. If the data were not normally distributed, we used a Mann–Whitney test (for two groups) or a Kruskal–Wallis H test followed by Dunn’s test (for multiple groups). The difference between comparisons was considered to be significant when P < 0.05. P values are indicated graphically: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. S1 shows validation of LRRK2-expressing SH-SY5Y cell lines and effects of ROC/COR and kinase-inactivating mutations on cytosolic Ca²⁺. Fig. S2 shows effects of disrupting lysosomes on cytosolic Ca²⁺ in SH-SY5Y cells. Fig. S3 shows effects of TPC2 agonists on cytosolic Ca²⁺ in SH-SY5Y cells. Fig. S4 shows effects of TPC2 agonists on cytosolic Ca²⁺ in human dopaminergic neurons. Fig. S5 shows effects of TPC2 and LRRK2 expression in Drosophila dopaminergic neurons. Table S1 lists iPSC lines used in this study. Table S2 lists primers used for quantitative PCR. Table S3 lists primers used for cloning and mutagenesis. Video 1 shows Ca²⁺ measurements in the Drosophila brain.

The data are available from the corresponding author upon reasonable request.

We thank Christopher Elliot (University of York) for initiating experiments in Drosophila and for his significant contributions to this project prior to his retirement. We also thank Alex Agrotis, Eric J. Lambie, and Yu (Cara) Yuan (all at UCL) for helpful comments, William Andrews (UCL) for molecular biology assistance, and Vil Fernandez (UCL) for her support with Drosophila imaging. Original observations were made by Bethan S. Kilpatrick and followed up by Adam Fitchett and Harry Ward (all at UCL).

This work was supported by Parkinson’s UK grants K-1802 (to S. Patel and A.H.V. Schapira) and G-2006 (to S. Patel, A.H.V. Schapira, and S.T. Sweeney); Wellcome Trust grants WT102129/B/13/Z (to M. Gregori and S. Patel), and WT088145AIA and WT095010MA (to K. Harvey); São Paulo Research Foundation (FAPESP) grant 2022/15748-0 (to G.J.S. Pereira); National Institutes of Health grant K08HL168147 (to X. Cai); Medical Research Council grants MR/M00676X/1 (to K. Harvey) and MR/T046007/1 (to A.H.V. Schapira); DFG grants GR4315/2-2, ANR/DFG, GR4315/6-1, SFB/TRR152 P04, SFB1328 A21, and GRK2338 P08 (to C. Grimm) and BR 1034/7-1 (to F. Bracher); and Department of Science and Technology, Govt of India SERB fellowship SB/DF/001/2019 (to G. Hasan). Open Access funding provided by University College London.

Author contributions: M. Gregori: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, validation, and visualization. G.J.S. Pereira: conceptualization, formal analysis, investigation, methodology, validation, visualization, and writing—review and editing. R. Allen: data curation, formal analysis, investigation, visualization, and writing—review & editing. N. West: conceptualization, data curation, and investigation. K.-Y. Chau: data curation, formal analysis, investigation, resources, visualization, and writing—review and editing. X. Cai: data curation, formal analysis, and visualization. M.P. Bostock: investigation. S.R. Bolsover: investigation. M. Keller: investigation and resources. C.-Y. Lee: investigation, methodology, resources, and supervision. S.H. Lei: resources. K. Harvey: conceptualization, funding acquisition, resources, supervision, and writing—review and editing. F. Bracher: funding acquisition, investigation, and writing—review and editing. C. Grimm: supervision and writing—review and editing. G. Hasan: resources. M.E. Gegg: investigation, methodology, and writing—review and editing. A.H.V. Schapira: conceptualization, funding acquisition, project administration, resources, supervision, and writing—review and editing. S.T. Sweeney: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, and writing—original draft, review, and editing. S. Patel: conceptualization, formal analysis, funding acquisition, methodology, project administration, supervision, visualization, and writing—original draft, review, and editing.