SNX10 at the crossroad of endocytosis and piecemeal mitophagy

ABSTRACT

Mitophagy targets damaged or dysfunctional mitochondria for lysosomal degradation. While canonical mitophagy pathways target the whole mitochondria for lysosomal degradation, it has become clear that selected mitochondrial components can be targeted for lysosomal degradation via other pathways, such as piecemeal mitophagy or mitochondria-derived vesicles. In a recent study, we identified the PX domain-containing endosomal protein SNX10 as a negative modulator of piecemeal mitophagy. Endosomal SNX10-positive vesicles dynamically interact with mitochondria and acquire selected mitochondrial proteins upon hypoxia. Zebrafish larvae lacking Snx10 show elevated Cox-IV degradation, increased levels of reactive oxygen species (ROS), and ROS-dependent neuronal death.

Mitochondria undergo constant quality control to maintain homeostasis. Mitophagy, a selective form of autophagy, maintains mitochondrial quality by targeting damaged or dysfunctional mitochondria for lysosomal degradation and recycling. While mitophagy has long been considered to facilitate the degradation of the whole mitochondria via PINK1/Parkin-dependent or independent pathways, increasing evidence suggests the existence of multiple mitochondrial degradation pathways that target selected parts of the mitochondria for lysosomal degradation, dependent on the core autophagic machinery or not. Recent research highlights dynamic interactions between mitochondria and endo-lysosomes as a novel aspect of mitochondrial quality control.

Given the growing complexity of mitochondrial degradation pathways, we aimed to identify novel regulators of hypoxia-induced mitophagy. Sorting Nexin 10 (SNX10) was identified as a negative modulator of mitophagy in a high-throughput targeted siRNA screen using hypoxia-mimicking conditions (deferiprone; DFP). SNX10 is the simplest member of the sorting nexin family, containing a phosphatidylinositol 3-phosphate (PI(3)P)-binding PX domain, and has primarily been associated with endosomal trafficking. In a recent paper [1], we demonstrate that SNX10 plays an important role at the crossroads of endocytosis and mitophagy. SNX10 was found to localize to both early and late endosomal compartments in a PI(3)P-dependent manner and to co-occur with internalized EGFR following EGF stimulation. Electron microscopy, western blotting, and widefield fluorescence microscopy revealed that cells lacking SNX10 had smaller EGFR-containing endosomes, slower EGFR degradation, and an increased number of EEA1-positive structures, indicating a role of SNX10 in the modulation of endosomal maturation. In line with this, a recent paper demonstrates that SNX10 regulates the conversion of PI(3)P to PI(3,5)P₂ in the endo-lysosomal pathway.

To further understand the cellular function of SNX10, we analyzed its interactome by mass spectrometry and identified associations with mitochondrial (COX-IV, ATP5J, ATPIF1), autophagic (SQSTM1/p62), and endosomal (MVB12A, LAMTOR1) proteins. Intriguingly, live-cell imaging of cells expressing SNX10-EGFP showed that SNX10-positive vesicles dynamically interact with mitochondria and occasionally incorporate mitochondrial components under DFP conditions. To explore the nature of mitochondrial cargo within SNX10-positive vesicles, we performed immunofluorescence analysis in SNX10-EGFP cells. Upon DFP treatment, ATP5J, SAMM50, and COX-IV were detected within SNX10-positive vesicles, while TOMM20, TIMM23, and PDH were rarely observed. In agreement with this, SNX10 depletion did not affect the formation of mitochondria-derived vesicles (MDVs) and had no significant effect on TOMM20 protein levels.

Interestingly, depletion of SNX10 significantly reduced COX-IV levels, as observed by western blot and immunofluorescence analysis. This was independent of COX-IV transcription or proteasomal degradation, reinforcing the role of SNX10 in the lysosomal processing of selected mitochondrial proteins. SNX10-positive vesicles containing COX-IV were also positive for LC3B and LAMP1, indicating a role for SNX10 in the degradation of mitochondrial material through the auto-lysosomal pathway. Pharmacological inhibition of VPS34 or ULK1, key regulators of macroautophagy, did however not prevent the increased COX-IV clearance observed in SNX10-depleted cells, indicating that SNX10 functions independently of the canonical macroautophagy pathway. It is not clear whether piecemeal mitophagy depends on ULK1 or VPS34, but it has been shown to rely on the autophagy receptor p62/SQSTM1. Indeed, SNX10-positive vesicles containing mitochondrial cargo co-localized with p62 and COX-IV levels increased in p62-depleted cells, suggesting a role for SNX10 as a negative modulator of piecemeal mitophagy (Figure 1).

Figure 1.

SNX10 modulates piecemeal mitophagy. SNX10 localizes to early endosomes via the binding of its PX domain to PI(3)P. Additionally, SNX10 localizes to late endosomes and was found to promote the trafficking of the EGFR in control conditions. Under hypoxia-mimicking conditions (using DFP or DMOG), SNX10-positive late endosomes co-occur with LC3 and incorporate selective mitochondrial components recognized by p62. SNX10 was found to inhibit the turnover of selected mitochondrial proteins, including COX-IV, indicating a role for SNX10 in piecemeal mitophagy. In vivo, zebrafish larvae lacking Snx10 (snx10ab double knock-out (DKO)) are characterized by increased levels of reactive oxygen species (ROS) and ROS-mediated cell death, as shown by increased TUNEL staining. Created in BioRender. Simonsen, A. (2025) https://BioRender.com/efesd4g.

Corroborating a role for SNX10 in the regulation of mitochondrial homeostasis, Seahorse XF analysis demonstrated a significant decline in mitochondrial respiration and ATP production in U2OS cells depleted of SNX10. We also observed a reduction in citrate synthase (CS) activity, although total CS protein levels remained unchanged, indicating impaired mitochondrial metabolism.

To assess whether Snx10 depletion in vivo had a similar effect on mitochondrial homeostasis, we used zebrafish as a model organism. Both duplicated gene copies of snx10 (snx10a and snx10b) were expressed at early developmental with particularly high expression in the brain as evident from qPCR and in-situ hybridization experiments, respectively. CRISPR/Cas9 depletion of both snx10a and snx10b demonstrated that zebrafish larvae lacking Snx10 exhibited significantly reduced levels of Cox-IV, confirming our in vitro data. Intriguingly, the level of reactive oxygen species (ROS) was significantly increased in snx10ab double knock-out (DKO) larvae, suggesting impairment in mitochondrial bioenergetics. TUNEL staining on snx10ab DKO larvae showed high neuronal cell death as compared to controls, which was reversed in larvae subjected to antioxidant treatment. Thus, our data indicate that increased mitophagy in larvae lacking Snx10 results in elevated oxidative stress and increased ROS-mediated neuronal cell death. Although mitophagy is widely considered a protective pathway, our data indicate that excessive mitophagy also can be detrimental. In line with this, larvae treated with DMOG, a hypoxia-mimetic mitophagy-inducing agent, showed increased neuronal cell death (Figure 1).

In conclusion, our data provide evidence that SNX10 plays a critical role in modulating the piecemeal mitophagy of specific OXPHOS machinery proteins, particularly under hypoxia-mimicking conditions when ATP is mainly produced through anaerobic glycolysis. This process enables cells to selectively degrade surplus mitochondrial components while preserving other mitochondrial components necessary for cellular function. Further work is now required to mechanistically characterize how SNX10 modulates piecemeal mitophagy of selected mitochondrial components. One possibility is that SNX10-mediated regulation of PI(3)P to PI(3,5)P2 conversion modulates the fusion of piecemeal mitophagosomes with the endolysosomal pathway. Another explanation is that SNX10 on endosomes acts as a scaffold that regulates the recruitment of the autophagic machinery to endosomal vesicles.

Funding Statement

This work was funded by the Norwegian Cancer Society [grants no. 171318 and 223278], from the European Union’s Horizon 2020 research and innovation program under the Marie Skłodowska-Curie [grant no. 801133], the South-Eastern Norway Regional Health Authority [grant no.2020032], and the Research Council of Norway through its Centers of Excellence funding scheme [grant no. 262652] and FRIPRO [grant no. 249753].

Disclosure statement

No potential conflict of interest was reported by the author(s).