Long-distance autophagy

Keywords:

• lysosome
• macroautophagy
• neuron
• phagophore
• stress

Much attention has been paid to elucidating the mechanism of phagophore nucleation and expansion, although many questions remain unanswered. For example, the origin of the membrane involved in these processes, the means by which curvature is generated, and the role of the ATG and other proteins in these steps are just a few of the pressing issues. An additional complication needs to be considered when dealing with neurons. Both dendrites and axons may project from the soma, and these extensions can be quite long—up to 1 m in humans. Although dendrites are shorter, they can still reach lengths close to 2 mm. Thus, neuronal cells, with their unique morphology, need to deal with the problem of autophagy at the extreme boundaries of the cell. How is this accomplished?

There are at least 3 distinct mechanisms employed by neurons to carry out macroautophagy with regard to where the process takes place. (1) Autophagosomes formed at the periphery can be transported along microtubules to the cell body where they fuse with lysosomes; however, these autophagosomes can also fuse with lysosomes during transport.^1,2,7 (2) The entire process of autophagy can be carried out at the periphery of the cell, such as in the distal axon.³ (3) The axon may rely on transcellular autophagy, as recently demonstrated with transmitophagy, to degrade the cargo in a neighboring cell.⁴

Long-range unidirectional transportation of autophagosomes along microtubules has long been observed in neurons.^5-7 Due to the challenge imposed by the complicatedly elongated nature of nerve cells, such retrograde transport of autophagosomes across axons necessitates a highly processive, spatially regulated mechanism of action. This process is also further complicated by the fact that axonal autophagosomes associate with both dynein and kinesin,⁷ cellular motors that facilitate retrograde and anterograde movement, respectively. If autophagosomes remain bound to both motors, how is a unidirectional movement toward the soma possible? How is the kinesin-mediated anterograde movement inhibited? In order to shed light into this highly complex process, Fu et al. investigated a possible role of the motor scaffolding protein MAPK8IP1/JIP1 in facilitating the retrograde transport of neuronal autophagosomes localized in axons.^1,2 Through a series of elegantly executed fluorescence microscopy assays coupled with knockdowns and mutational analyses, the authors showed that MAPK8IP1 employs a multi-layered mechanism to facilitate the robust retrograde movement of autophagosomes along an axon. Initially, autophagosomes formed in distal axons bind both dynein and kinesin and thus, undergo unregulated, bidirectional movement. However, once MAPK8IP1 is recruited to LC3-coated autophagosomes via its LIR motif, it loses its ability to bind kinesin's heavy chain (KIF5). This effectively inhibits anterograde motion allowing autophagosomes to exit the distal end of the axon and move more processively toward the soma.

An additional layer of control is also utilized by MAPK8IP1 by means of post-translational modification. Phosphorylation of MAPK8IP1 at residue S421 serves as a switch that turns on binding to KIF5, which in turn results in activation of anterograde motion. Indeed, knocking down endogenous MAPK8IP1 and subsequent replacement with a phosphomimetic version supports this functional switch hypothesis; neurons containing MAPK8IP1 S421D exhibit not only a reduction in retrograde movement of autophagosomes, but also a marked increase in the number of immobile, bidirectional, and forward-moving autophagosomes. Corroboratively, when MAPK8IP1 S421 is mutated into an alanine, retrograde movement is rescued. Overall, the work carried out by Fu et al.,^1,2 lays out an elegant mechanism by which autophagosomes achieve unidirectional motion along an axon culminating in the fusion of the autophagosome with a lysosome, allowing its contents to be degraded and recycled.

The retrograde transport of autophagosomes across millimeters of axons illustrates how neurons have overcome spatial limitations in order to clear up unwanted proteins and organelles at a basal level in locations distal from the cell body. However, when time is of the essence (i.e., highly toxic, damaged organelles need to be quickly engulfed and efficiently degraded), Ashrafi et al.^3,8 show that autophagy can occur in outlying regions of a neuron, eliminating the need for canonical long-distance hauling of autophagosomes. In this publication, the authors utilize 2 methods to site-specifically induce mitochondrial damage to determine whether or not mitophagy could actually transpire in axons. The first method involves the use of a genetically encoded, mitochondria-targeted, light-inducible photosensitizing agent, mt-KR.^9,10 Upon exposure to 555-nm light, mt-KR triggers damage to approximately 3-4 mitochondria at a time within a 10-μm section of an axon. In the second method, the authors employed microfluidics to isolate hippocampal neurons and locally treat them with antimycin A. The authors showed that activated mt-KR and antimycin A drastically damage the mitochondria resulting in fragmentation, swelling, and depolarization. Furthermore, these damaged axonal mitochondria stay immobile, colocalizing with autophagosomes. These stationary, mitochondria-containing autophagosomes are subsequently engulfed by axonal lysosomes as evidenced by the colocalization of RFP-LC3 and LAMP1-YFP. The authors then asked which cellular factors sense damaged mitochondria and stimulate the induction of mitophagy in axons. To this end, Ashrafi et al.³ looked at the usual suspects, PARK2 and PINK1, both of which facilitate mitophagy. The authors examined whether or not these proteins are also involved in the axonal elimination of faulty mitochondria. Indeed, after selectively inducing damage to distal mitochondria, PARK2 is rapidly recruited. Furthermore, when Park2 or Pink1 is deleted, mitophagy is not observed in axons. In summary, the findings in this paper demonstrate that localized mito-phagy can occur in regions distal from a neuron's soma. These observations are novel and clearly illustrate an example of how cells, in particular, neurons, evolve mechanisms to protect themselves from oxidative stress and damage, ridding the cells of toxic substances quickly and efficiently.

In another recent publication examining mitophagy in neurons, Davis et al.^4,11 report a previously undiscovered phenomenon they termed transmito-phagy: the trans-cellular degradation of axonal mitochondria by an adjacent cell. This observation provides another attractive model as to how damaged mitochondria can be rapidly removed from axons without the need for retrograde transport. This mechanism gives neurons a temporal advantage in effectively preventing further neuronal damage that can be caused by reactive oxygen species. Using serial block-face scanning electron microscopy, the authors presented elegant pictures of clustered mitochondria within retinal ganglion cell axons at locations adjacent to astrocytes. These huddled mitochondria form protrusions that are eventually separated from axons upon engulfment by a neighboring astrocyte. The authors also tracked the degradation of these mitochondria through the use of tandem fluorescent protein markers: pH sensitive EGFP, and pH-impervious mCherry. These 2 agents were fused to the mitochondrial resident protein COX8A (cytochrome c oxidase subunit VIIIA [ubiquitous]) and delivered to the retina. Solitary mCherry puncta are observed only when the mitochondria become associated with the lysosome (as indicated by LAMP1) and when they colocalize with the cytoplasm of an astrocyte (as indicated by LGALS3/MAC-2). Finally, using a combination of MitoFISH and TUNEL, the authors followed the fate of axonal mitochondrial DNA and observed that large amounts of these labeled nucleic acids are found within astrocytes.

The complex shape of a neuron poses many significant challenges as to how intercellular communication coordinating multiple cellular processes takes place in this type of cell. The study of autophagy in neurons has provided us a glimpse as to how these cells have cleverly evolved molecular machineries to overcome the temporal and spatial dilemma imposed by unique morphology. Many questions, however, still remain. How do autophagosomes decide whether or not to undergo retrograde transport or remain in distal axons for lysosome fusion? How are PARK2 and PINK1 transported to distal axons leading to their subsequent recruitment to dysfunctional mitochondria? How does signaling occur across 2 different cells to facilitate transmitophagy? Without a doubt, further investigations into this matter will uncover many more distinct and elegant mechanisms giving us a clearer understanding of how autophagy transpires in neuronal cells.

Acknowledgment

The authors thank Dr. Erika Holzbaur (University of Pennsylvania) for helpful comments.

Funding

This work was supported by NIH grant GM053396 (to DJK).