This study shows that 10-day inhibition of glucocerebrosidase activity in a dopaminergic cell line could modulate susceptibility to ferroptosis by modifying lipid composition and altering protein degradation of key players like GPX4 involved in the regulation of ferroptosis. However, this phenomenon is transient, suggesting that the cells have the capacity to return to a state of basal homeostasis, highlighting the complexity of cellular mechanisms.
First identified as an original cell death pathway, an increasing number of studies indicate that ferroptosis may occur by sharing common effectors with other types of regulated cell death like autophagy24,25 and lysosomal cell death.26,27 Mutations in GBA inducing a loss of function of a lysosomal enzyme have been associated with a higher risk of developing PD. Thus, we studied the interaction between lysosomal storage disorder induced by glucocerebrosidase inhibition and ferroptosis in dopaminergic neurons.
CBE is a well-known and effective inhibitor of lysosomal glucocerebrosidase.28 Given that inhibition of a single lysosomal enzyme takes time to alter a system such as the lysosomal pathway, the inhibition of glucocerebrosidase for several days appeared more relevant in LUHMES cells. Our study shows that up to 20 days of treatment, CBE did not impact LUHMES cell viability in basal conditions. CBE treatment inhibited lysosomal glucocerebrosidase activity without impacting its mRNA or protein levels and increased glycosphingolipid species. Thus, long-term impairment of glucocerebrosidase activity can be modelled in this dopaminergic neuron model. Long-term inhibition of glucocerebrosidase activity using CBE, for 14 or 29 days, has already been achieved in another dopaminergic neuronal model derived from human induced pluripotent stem cells.29
LUHMES cells are sensitive to many toxins allowing the study of several cell death pathways. Short- or long-term glucocerebrosidase deficiency in LUHMES cells did not alter sensitivity to several cell death inducers. CBE cells were protected against rapamycin, probably due to disruption of lysosomal trafficking. More importantly, 10-day CBE cells were transiently and strongly protected against RSL3-induced ferroptosis while no protection against erastin was observed. Erastin and RSL3 promote the reduction of antioxidant activity by inhibiting the cystine-glutamate transmembrane antiporter causing cysteine and glutathione depletion or by directly inactivating GPX4, respectively. In the SOD1 transgenic mouse model of amyotrophic lateral sclerosis, a study showed that CBE preserves motor neurons.30
To discard an artefactual interaction between CBE and RSL3, another pro-ferroptotic condition consisting of a combination of AA and ferric chloride was used. It has previously been demonstrated that by enriching the membrane with AA, LUHMES cells are more sensitive to GPX4-dependent lipid peroxidation.18 10-day CBE cells were protected against this co-treatment while no difference was observed after 20 days. Oxidative stress analysis supported these results since a significant reduction in RSL3-induced lipid peroxidation was observed in 10-day CBE cells.
As glucocerebrosidase inhibition seemed to specifically affect lipid composition and peroxidation, we assessed the expression of the main players involved in the regulation of ferroptosis. A transient upregulation of GPX4, HMOX1, and FTH1 proteins levels was observed in 10-day CBE cells but not 20-day CBE cells compared to controls. In contrast to these results, excessive activation of HMOX1 has been shown to enhance ferroptosis while pharmacological inhibition or silencing of HMOX1 confers resistance to ferroptosis.32–34 However, HMOX1 may also act in a cytoprotective manner, probably depending on the level of activation. The protective effect of HMOX1 is attributed to its antioxidant activity whereas its toxic effect is due to increased generation of ferrous iron, which stimulates the Fenton reaction. Thus, over-regulation of HMOX1 could be cytotoxic, while moderate regulation could be cytoprotective.35,36 Iron release and FTH1 degradation have already been shown to be regulated through ferritinophagy, a selective form of autophagy, which can release sufficient iron to trigger ferroptosis. In PC12 cells, FTH1 overexpression impairs ferritinophagy, suppressing ferroptosis-induced cell death.37 Furthermore, inhibition of lysosomal activity or silencing of nuclear receptor coactivator 4 (a cargo receptor recruiting FTH1 to autophagosomes for lysosomal degradation) suppresses ferroptosis.24,25,27 Interestingly, erastin but not RSL3 seems to promote ferritinophagy in HeLa cells.38 Therefore, CBE could be more efficient against RSL3 since the mechanism of action of RSL3 appears to be independent of the ferritin degradation process. Although the direct inactivation of GPX4 by RSL3 is uncertain,39 the role of GPX4 is central in scavenging lipid peroxidation. Subsequent to GBA inhibition, GPX4 is more abundant in the cells and its higher activity translated into an increased GSH/GSSG ratio. This is certainly one of the main effects of CBE that makes cells more resistant to RSL3-induced ferroptosis. As with FTH1, GPX4 regulation by autophagy is well documented.40 For example, acid sphingomyelinase, a key enzyme in sphingolipid metabolism, mediates activation of autophagy and induces GPX4 degradation. Conversely, inhibition of this enzyme reduces GPX4 degradation by autophagy.41
We then determined whether protein upregulation following GBA inhibition resulted from a defect in their degradation by the lysosomal pathway. In 10-day CBE cells, a disruption of the lysosomal pathway was measured by an increase in the LC3-II/LC3-I ratio, an increase in the number of lysosomes, and no difference in lysosome biogenesis. Impaired lysosomal activity was observed after long-term treatment with CBE in another dopaminergic neurons model. This lysosomal alteration was measured by increased levels of lysosomal associated membrane protein 1 and TFEB nuclear translocation as well as the number and size of lysosomes.29 A reduction in lysosome catalytic activity was also found in this model using radioactive glycosphingolipids.29 These results are consistent with our observations on lysosomal pathway impairment. To complete the demonstration that impaired autophagic flux may be responsible for resistance to ferroptosis, potent autophagic flux inhibitors were used. Short treatment was sufficient to perturb autophagic flux and protect against ferroptosis. It was noticeable that the inhibitor with the strongest effect on neuroprotection and lipid peroxidation, BafA1, seemed to have the least effect on GPX4 protein levels; a significant increase in GPX4 protein was observed for the HCQ and NH4Cl conditions. Finally, GPX4 silencing restored RSL3-induced ferroptosis sensitivity in 10-day CBE cells showing that the neuroprotection observed could result directly from increased GPX4 expression.
In recent years, the number of studies on ferroptosis has increased dramatically, revealing a surprising degree of complexity. Concerning the interaction between cell death pathways, it is a real challenge to define the threshold or checkpoints associated with pro-survival and pro-death autophagy in ferroptosis. Loss of GBA function has been associated with a higher risk of developing PD, probably due to long-term consequences on glycosphingolipids and lysosomal activity. This study shows that transient inhibition of GBA activity could have a positive effect on lipid peroxidation. If sustained for a relatively short length of time, by altering the lysosomal pathway, an inhibitor like CBE modified lipid composition and increased GPX4 stability. Various small molecule compounds can affect GPX4 expression, mainly leading to its degradation.42 It is therefore important to better understand the different pathways leading to GPX4 degradation in order to develop specific GPX4 activators to protect neurons against excessive lipid peroxidation and subsequent ferroptosis.