Mutations in the ATP13A2
gene cause KRS and early-onset parkinsonism (4
). The normal function of ATP13A2 and the pathogenic mechanism(s) by which recessive mutations precipitate neurodegeneration are poorly understood. Here, we demonstrate that endogenous ATP13A2 is enriched in the microsomal fraction from mouse brain, and exogenous ATP13A2 localizes to intracellular acidic vesicles within mammalian neurons, particularly lysosomes and endosomes, and to a smaller extent autophagosomes. We further demonstrate that ATP13A2 protein is localized to neuronal populations of the human cerebral cortex and substantia nigra where this protein is increased in PD and DLB brains. The overexpression or silencing of ATP13A2 compromises the integrity of midbrain dopaminergic neurons, leading to reduced neurite length but without evidence of overt cell death. Modulation of ATP13A2 expression in cortical neurons modestly alters the size and number of LC3-positive autophagosomes but without an obvious effect on autophagic activation. Consistent with a potential role in regulating vesicular cation transport, we find that silencing of endogenous ATP13A2 in cortical neurons modulates the kinetics of intracellular pH following cadmium-induced acidification and reduces intracellular calcium levels. Overexpression of human ATP13A2 similarly reduces intracellular calcium levels but without influencing pH. Finally, we show that silencing of ATP13A2 in neurons induces mitochondrial fragmentation, whereas ATP13A2 overexpression delays mitochondrial fragmentation induced by cadmium exposure. Collectively, our study reveals a number of intriguing phenotypes owing to loss- or gain-of-function of ATP13A2 and suggests a role for ATP13A2 in regulating vesicular cation transport and neuronal integrity. Furthermore, our data potentially implicate a role for ATP13A2 in idiopathic PD.
Using a specific polyclonal antibody to ATP13A2, we demonstrate the enrichment of ATP13A2 in the microsomal fraction from mouse brain which contains lysosomes, the Golgi complex and ER. We further demonstrate the localization of exogenous ATP13A2 to intracellular vesicular compartments within cortical neurons, particularly lysosomes, early and late endosomes, microtubule-associated vesicles and to a smaller extent autophagosomes. Previous studies examining exogenous ATP13A2 have demonstrated its localization largely to lysosomes in mammalian cell lines (6
). Therefore, the function of ATP13A2 may be broader than first anticipated with a preference for vesicular compartments that are known to be acidic. This might suggest a role for ATP13A2 in regulating the acidity of such vesicles, similar to the vesicular vacuolar-type H+
-ATPase, through the active transport of cations across vesicular membranes in an ATP-dependent manner. In yeast, an ortholog of ATP13A2 termed Ykp9 is localized to the vacuole where it is implicated in regulating the transport of heavy metal ions, including cadmium, nickel, manganese and selenium (16
). Of note, however, human ATP13A2 is unable to complement the phenotype of ykp9
null yeast, indicating a lack of functional conservation between Ykp9 and ATP13A2 (16
). The transport of heavy metal ions by mammalian ATP13A2 has not yet been directly demonstrated, and consequently, the cation selectivity of this transporter is not known. A recent study has shown that the overexpression of human ATP13A2 regulates intracellular manganese concentration and protects from manganese-induced toxicity in mammalian cell lines (18
). Whether or not these effects relate to the direct ATP-driven transport of manganese ions across membranes by ATP13A2 remains to be determined, and whether ATP13A2 can similarly regulate the homeostasis of other heavy metal ions is not clear.
In the human brain, we demonstrate the localization of ATP13A2 to various neuronal populations, including pyramidal neurons throughout the cerebral cortex and dopaminergic neurons of the substantia nigra. Furthermore, we detect ATP13A2 protein migrating as an ~130 kDa species in extracts of human striatum and medial frontal gyrus. A previous study has reported the expression of ATP13A2 mRNA throughout the human and mouse brain where it appears to adopt a neuronal localization with enrichment in the ventral midbrain and expression in individual dopaminergic neurons (6
). We can confirm that ATP13A2 protein is similarly localized to neurons, including nigral dopaminergic neurons, throughout the mouse (data not shown) and human brain. We further demonstrate that ATP13A2 protein levels are increased in surviving substantia nigra dopaminergic neurons and cortical pyramidal neurons of PD brains compared with normal control brains. These findings are similar to a previous study which estimated a 10-fold increase in ATP13A2 mRNA in individual nigral dopaminergic neurons from PD brains compared with control brains (6
). The smaller increase in protein levels compared with mRNA levels of ATP13A2 in PD brains could relate to inefficient translation of mRNA, reduced mRNA stability or to differences in post-mortem brain tissue. Nevertheless, these data collectively indicate that ATP13A2 protein levels are increased in vulnerable neurons from the substantia nigra, striatum and cerebral cortex of PD brains. The reason for this increase is unclear, but could relate to a compensatory neuroprotective response in neurons exposed to PD and/or DLB pathogenic processes that potentially correlates with up-regulation of the autophagy–endosomal–lysosomal pathways. We have previously shown that the autophagosomal marker LC3 is increased in neurons in vulnerable areas of DLB brains (29
). Moreover, we demonstrate that ATP13A2 protein is commonly detected in neurons harboring LB pathology and that the levels of cytoplasmic ATP13A2 are increased in LB-positive neurons compared with neurons without LBs. While ATP13A2 does not co-localize with LBs directly, this might suggest a compensatory up-regulation of ATP13A2 or acidic vesicles in response to LBs. It is unlikely that increased ATP13A2 levels directly impair neuronal viability since (i) this protein is increased in ‘surviving’ nigral dopaminergic neurons of PD brains, (ii) the overexpression of ATP13A2 is reported to protect from neuronal toxicity induced by α-synuclein and manganese (16
) and (iii) loss of function of ATP13A2 due to recessive mutations induces neurodegeneration (4
), consistent with a normal role for ATP13A2 in neuronal maintenance and/or neuroprotection in humans.
In attempting to model the pathogenic effects of recessive, loss-of-function ATP13A2
mutations in primary midbrain dopaminergic neurons, we find that silencing of endogenous ATP13A2 expression reduces neurite outgrowth but without causing neuronal cell death. This effect is somewhat selective since silencing of ATP13A2 in primary cortical neurons does not impair neurite outgrowth or induce cell death. Surprisingly, the overexpression of human ATP13A2 also markedly reduces neurite outgrowth of dopaminergic neurons. Therefore, it appears that there is a critical range of ATP13A2 levels required for the proper maintenance of dopaminergic neuronal processes. Conceivably, ATP13A2-mediated cation transport across vesicular membranes is a tightly regulated process, and dysregulation of this process could potentially impact the normal function of acidic vesicles, axonal transport and the subsequent maintenance of neuronal processes. Recent studies have shown that the overexpression of ATP13A2 orthologs protects from toxicity induced by mutant α-synuclein in yeast, worm and cultured primary dopaminergic neuronal models (16
). Although we have not been able to show a detrimental effect of ATP13A2 overexpression on neuronal viability, we find instead that the integrity of neuronal processes is compromised. Therefore, a potential note of caution is warranted for neuroprotective strategies that seek to increase ATP13A2 levels or activity in pre-clinical models of PD. It is interesting to note that the lentiviral-mediated overexpression of ATP13A2 in cultured dopaminergic neurons appears to cause mild neuronal loss (16
). However, in our hands, using single-cell analysis, we do not observe obvious neuronal loss or apoptotic cell death. The functional implications of reduced neurite length induced by silencing or overexpression of ATP13A2 in dopaminergic neurons are unclear at present and will await the future development and analysis of knockout or knockdown mice. Similarly, the reported protective effects of ATP13A2 overexpression against α-synuclein-induced toxicity will require confirmation in animal models of PD (16
Modulation of ATP13A2 expression in cultured primary cortical neurons produces a range of intracellular phenotypes. ATP13A2 appears to modestly influence the number and size of LC3-positive autophagosomes, but does not appreciably alter autophagic activation based on the levels of the autophagosomal marker, LC3-II. Similar experiments conducted in human neural cell lines did not reveal a consistent effect of modulating ATP13A2 expression on autophagic activation or flux (data not shown). Instead, in cortical neurons, we observe a reduction in steady-state total GFP-LC3 levels following ATP13A2 silencing or overexpression which may indicate increased lysosomal turnover. We also demonstrate that ATP13A2 silencing modestly increases the rate of intracellular acidification induced by acute cadmium exposure but without an effect on basal pH. These data may suggest that lysosomes and related acidic vesicles exhibit less tolerance to alterations in pH following ATP13A2 silencing, indicating that this protein may play a selective role in maintaining the homeostasis of H+
ions under conditions of cellular stress. It is not clear whether vesicular pH is similarly affected by ATP13A2 silencing. Yeast cells lacking ykp9
appear most sensitive to growth on media containing cadmium when compared with other heavy metals (17
). Our data tend to support a functional connection between ATP13A2 and cadmium, whereas manganese exposure failed to influence intracellular pH in these assays, thus preventing a similar functional assessment.
We also find that both silencing and overexpression of ATP13A2 reduce the basal levels of intracellular calcium ions, an effect that is maintained following cadmium-induced calcium release. While these observations do not support a direct relationship between ATP13A2 and cadmium in these assays, they instead suggest that ATP13A2 may play a role in regulating calcium homeostasis. Whether or not this effect is through direct ATP13A2-mediated transport and sequestration of calcium ions into acidic vesicles is unclear at present, but it could also result from a general impairment of cation homeostasis in neurons which might indirectly influence calcium storage and/or availability. The ability to release intracellular calcium induced by cadmium exposure is still intact in neurons following ATP13A2 silencing, suggesting a modest but not essential role for ATP13A2 in the sequestration of calcium into intracellular stores. Such a role for ATP13A2 may potentially relate to an indirect effect on the general ionic homeostasis of neurons due to the dysregulation of acidic vesicles. ATP13A4, a related member of the P5
-type ATPase subfamily, is localized to the ER where its overexpression in COS-7 cells increases intracellular calcium levels (25
). Therefore, two P5
-type ATPases are potentially involved in calcium regulation albeit within distinct subcellular locations, i.e. ATP13A4 in the ER and ATP13A2 in lysosomes and related acidic vesicles. It will be interesting to determine whether additional P5
-type ATPases are also involved in regulating intracellular calcium levels. We further show that ATP13A2 influences mitochondrial morphology potentially as a precursor to mitochondrial damage. Similar to the effects on intracellular calcium, silencing of ATP13A2 expression induces basal mitochondrial fragmentation and increases the rate of mitochondrial fragmentation induced by cadmium exposure. The overexpression of ATP13A2 did not influence mitochondrial morphology, but instead delayed cadmium-induced mitochondrial fragmentation consistent with a neuroprotective effect in this paradigm. ATP13A2 harboring an F182L mutation, identified as a homozygous variant in a Japanese subject with early-onset parkinsonism (13
), failed to similarly delay mitochondrial fragmentation, suggesting that this disease variant causes a loss of function. The effects of modulating ATP13A2 expression on mitochondrial fragmentation do not initially appear to correlate with their effects on intracellular calcium levels and pH. Furthermore, ATP13A2 fails to co-localize with mitochondria in cortical neurons (data not shown), suggesting an indirect effect on mitochondria potentially through altered vesicular cation transport and cation homeostasis.
Our data clearly demonstrate that silencing of ATP13A2 expression, which is anticipated to model recessive disease, is consistently detrimental to neurons by manifesting dopaminergic neurite shortening and by altering the regulation of intracellular pH, reducing intracellular calcium levels and inducing mitochondrial fragmentation in primary cortical neurons. Therefore, neuronal phenotypes due to ATP13A2 silencing may provide potential insight into the molecular pathways underlying neurodegeneration due to recessive ATP13A2
mutations in KRS and early-onset parkinsonism. However, for reasons that are not yet clear, the overexpression of human ATP13A2 produces varied phenotypes including dopaminergic neurite shortening and reduced intracellular calcium levels, protection from cadmium-induced mitochondrial fragmentation, but no effect on intracellular pH regulation. Thus, ATP13A2 overexpression appears to produce both detrimental and protective effects on neurons, depending on the specific phenotype assessed. It will be important to determine in future experiments whether or not ATP13A2 overexpression produces an overall beneficial neuroprotective effect in cultures and in vivo
, consistent with previous studies reporting a protective effect against cytotoxicity induced by α-synuclein overexpression and manganese exposure (16
). Collectively, our data support a role for ATP13A2 in the maintenance of neuronal integrity and the regulation of intracellular cation homeostasis, autophagy and mitochondrial health. These broad cellular and subcellular effects may relate to the localization of ATP13A2 to intracellular vesicular compartments within neurons. Silencing of ATP13A2 in primary neurons may provide a useful model for understanding the normal function of this protein and for elucidating the molecular mechanisms underlying neurodegeneration due to recessive ATP13A2