Impaired Mitochondrial Mobility in Charcot-Marie-Tooth Disease

Charcot-Marie-Tooth (CMT) disease is a progressive, peripheral neuropathy and the most commonly inherited neurological disorder. Clinical manifestations of CMT mutations are typically limited to peripheral neurons, the longest cells in the body. Currently, mutations in at least 80 different genes are associated with CMT and new mutations are regularly being discovered. A large portion of the proteins mutated in axonal CMT have documented roles in mitochondrial mobility, suggesting that organelle trafficking defects may be a common underlying disease mechanism. This review will focus on the potential role of altered mitochondrial mobility in the pathogenesis of axonal CMT, highlighting the conceptional challenges and potential experimental and therapeutic opportunities presented by this “impaired mobility” model of the disease.


INTRODUCTION
Charcot-Marie-Tooth (CMT) disease is the most commonly inherited neurological disorder, affecting ∼1 in 5000 people (Skre, 1974;Barreto et al., 2016). It is a peripheral neuropathy defined by progressive deterioration of the peripheral nerves in the distal parts of the body, specifically the feet, hands, and lower extremities. This typically results in both motor and sensory loss in the affected areas. Unraveling the pathogenic mechanism(s) underlying CMT is somewhat complicated by the fact that CMT is both genetically and clinically heterogeneous. CMT has many subtypes including, demyelinating (affecting mainly Schwann cells), axonal, and intermediate (affecting both axons and Schwann cells). Herein, we are focusing on perturbations of mitochondrial mobility that might underlie the pathogenesis of axonal CMT.
Charcot-Marie-Tooth variants were originally classified based purely on clinical data. However, a recent explosion in genetic data can be mined to generate some compelling hypotheses. To date, over 100 mutations across more than 40 different proteins have been implicated in axonal and intermediate CMT. A large fraction of CMT-associated proteins has been shown or is predicted to affect the mobility of mitochondria or other organelles ( Table 1). In this review, we will focus on those that affect mitochondrial mobility and hypothesize that defects in this process might begin to explain why CMT mutations mainly affect peripheral neurons. At the same time, we highlight important limitations to this "impaired mobility" model. Our belief is that the insights gained from studying the effects of CMT mutations in peripheral neurons will inform the role of mitochondrial mobility in other types of neurons and neurodegenerative disorders, including those associated with aging.

CMT IS A PROGRESSIVE DISORDER THAT AFFECTS PREDOMINANTLY THE LONGEST NEURONS
Charcot-Marie-Tooth usually affects only the feet, hands, and lower extremities. The axons leading to these distal sites can be as long as a meter in some individuals. There have been some reports of central nervous system involvement but these instances are rare (Pareyson and Marchesi, 2009;Lee et al., 2017). Some CMT mutations also cause optic atrophy, and the optical nerve notably consists of relatively long axons (∼50 mm). Patients are usually born unaffected, but typically by age 10 display major losses of function. However, the range of age can be from the toddler years to the 5th decade of life (Skre, 1974;Verhoeven et al., 2003;Zuchner et al., 2004Zuchner et al., , 2006Chung et al., 2006;Engelfried et al., 2006;Cho et al., 2007;Calvo et al., 2009;Braathen et al., 2010;Boyer et al., 2011b). The severity of the disease is directly correlated with the age of onset (Chung et al., 2006;Verhoeven et al., 2006), and the longer axons (i.e., the feet) invariably degenerate before the shorter axons (i.e., the hands). In cases where CMT patients also display optic atrophy, this occurs after loss of function in the hands (Verhoeven et al., 2006;Zuchner et al., 2006). Together, these observations indicate a disease that directly correlates the length of the axon with the speed of onset, and the speed of onset with the magnitude of the pathology.

CMT MUTATIONS LARGELY AFFECT MITOCHONDRIAL MOBILITY
As mentioned above, a unique peripheral nerve characteristic is their extreme length, which suggests these cells are uniquely sensitive to impaired long-distance transport. Put simply, a decrease in mobility would have a greater impact on longer distance commutes than shorter ones. In support of this theory, 23 out of the 48 genes mutated in axonal or intermediate CMT encode proteins that play roles in mitochondrial function, often impacting mitochondrial mobility ( Table 1).
How alterations in mitochondrial motility impact mitochondrial function, particularly in the context of CMT, remains poorly understood. Despite clear defects in mitochondrial mobility, some studies have concluded CMT mutations do not alter readouts of mitochondrial OXPHOS function such as mitochondrial membrane potential, oxygen consumption, and ATP production, or impair cellular calcium levels which mitochondria are involved in controlling (Baloh et al., 2007;Larrea et al., 2019). However, other studies have demonstrated that CMT mutations cause defects in these readouts (Loiseau et al., 2007;Guillet et al., 2010;Barneo-Munoz et al., 2015;Saporta et al., 2015;Rocha et al., 2018;Almutawa et al., 2019;Bernard-Marissal et al., 2019). And, another study demonstrated that bioenergetic efficiency and viability in a fly model can be rescued with only minor alterations in mitochondrial distribution (Trevisan et al., 2018). These discrepancies may be at least partially explained by differences in the model systems and experimental conditions used. There are now a wide variety of tools to study CMT including mouse and fly (Drosophila melanogaster) genetic models and iPSC-derived motor neurons (Saporta et al., 2015;Yamaguchi and Takashima, 2018;Juneja et al., 2019).
Recently, a screen for RAB7A binding partners found that another CMT protein, INF2 (Inverted Formin 2), is one of several actin-binding candidate interaction partners for RAB7A (Pan et al., 2020). This is particularly relevant to our discussion on CMT, inter-organelle contacts, and mitochondrial mobility for multiple reasons. First, a splice isoform of INF2 is tailanchored to the ER. Second, dominant active mutations in ERanchored INF2 that mimic INF2 CMT mutations have been shown to increase actin-dependent mitochondrial fragmentation and decrease mitochondrial mobility (Korobova et al., 2013;Chakrabarti et al., 2018). Together, these data point towards an important role in ER-mitochondria inter-organelle contacts in somehow regulating mitochondrial mobility via the actin cytoskeleton. That INF2 also potentially interacts with RAB7A suggests that mitochondria, endo-lysosomes, and ER all contact one another via CMT-associated proteins.
All INF2 CMT mutations are predicted or have been shown to increase actin assembly (Bayraktar et al., 2020). While some actin-binding motor proteins likely facilitate microtubule-independent mitochondrial transport, numerous studies have shown that long-range microtubule-based mobility of mitochondria is antagonized by actin and actin-binding motor proteins (Chada and Hollenbeck, 2004;Quintero et al., 2009;Pathak et al., 2010;Venkatesh et al., 2019;Cardanho-Ramos et al., 2020). Thus, while the effects of INF2 CMT mutations have yet to be studied in neurons, it is reasonable to expect that INF2 CMT mutations will cause an actin-dependent decrease in mitochondrial mobility in axons. Furthermore, since the ER regularly contacts many other organelles, and even appears to drive actin-assembly at ER-organelle contact sites (Korobova et al., 2013(Korobova et al., , 2014Manor et al., 2015;Chakrabarti et al., 2018;Yang and Svitkina, 2019;Schiavon et al., 2020), it is quite possible INF2 CMT mutations cause aberrant actin assembly on other organelles, reducing their mobility as well (Figure 1).
Together with the newly uncovered role for RAB7A in (indirectly) modulating actin assembly (Pan et al., 2020), these observations point towards a role for multiple CMT mutations causing aberrant organelle-organelle and organelle-actin contacts, all of which cause decreased mitochondrial mobility. Whether some (or all) CMT mutations also cause decreased mobility of other organelles remains an important open question.
The focus of the role of mitochondria in CMT has been primarily on MFN2 and GDAP1, and to a lesser extent on associated motor proteins (KIF1B -Kinesin Family Member 1B, KIF5A -Kinesin Family Member 5A, DYNC1H1 -Dynein Cytoplasmic 1 Heavy Chain 1, DCTN2 -Dynactin Subunit 2) and some cytoskeletal proteins (NEFL -Neurofilament Light). Here, we have highlighted INF2 and RAB7A as CMT-associated proteins likely involved in mitochondrial mobility and dynamics. However, we propose that the proteins mutated in CMT that play roles in mitochondrial function, dynamics and mobility likely extend well beyond just these two (see Table 1 for a full list).

WHY DO MOBILITY DEFECTS USUALLY ONLY AFFECT PERIPHERAL NEURONS IN CMT PATIENTS?
Hopefully, we have provided a convincing argument that many CMT mutations likely reduce mitochondrial mobility. Given the extreme lengths of peripheral axons, it is tempting to conclude that a reduction in mobility due to CMT mutations simply affects longer axons more severely (the "impaired mobility model" of CMT). One can easily reconcile two key features of CMT using the impaired mobility model: The progressive nature of the disease: This suggests that dysfunction must accumulate over time. One can imagine this more severely affects longer axons, due to reduced turnover of damaged mitochondria resulting from reduced mitochondrial mobility. Interestingly, one could imagine that reduced mobility of other organelles associated with turnover (e.g., lysosomes) could also cause increased accumulation of damaged mitochondria in longer axons. The longest peripheral axons (i.e., the feet) progressively degenerate prior to the next-longest axons (i.e., the hands): This further supports the impaired mobility model, wherein damage accumulates first in the longest axons due to the more demanding, "longer commute" resulting in faster accumulation of damaged mitochondria.
Unfortunately, while this model is compelling, it appears to be overly simplistic. The weakness in relying on mobility alone as an explanation can best be highlighted by comparing the lengths of different axons both within and between species. For example, the longest axon in mice is approximately 2 cm, whereas in humans the longest axon is ∼60 times longer. Just as striking, some unaffected axons in the human CNS may be longer than the mouse's longest axon. The very same mutation in humans and mice can cause CMT, yet no defects are found in the brains of human CMT patients. Meanwhile, the motor proteins, cytoskeletal tracks, and mitochondria of mice and humans are all roughly the same size, and all possess roughly the same biophysical properties (e.g., velocity, force generation, etc.) when transporting their organelles across long distances (Figure 2). FIGURE 2 | The relative scales of axons in humans versus mice. While microtubules, motor proteins, and mitochondria have nearly identical sizes in mice and men, the length of the axons can be very different. At the same time, axons in the mouse sciatic nerve are not as long as some of the longer axons within the human brain. For example, it is known in the macaque there are direct connections from the frontal eye fields in the anterior bank of the arcuate sulcus to the primary visual cortex, which in the macaque is ∼5 cm, and probably as long as 12.5 cm in humans. Notably, for most CMT patients the pathology is often constrained to only peripheral neurons. Since the same mutations can cause CMT in mice and men, the distances mitochondria must travel can only provide a partial explanation for the physiopathology of these mutations.
Any analysis of mobility must consider not just distance but also time. Most laboratory mice only live for ∼2 years, while human CMT patients may not even experience symptoms until adolescence or adulthood. Thus, there is clearly a "missing variable" that underlies differences in lifespan and disease susceptibility between species (e.g., differences in metabolism or oxidative stress). Thus, CMT may serve as a "model disease" to better understand age-related neurodegeneration. It is well established mitochondria play myriad roles at the presynapse, including ATP production, intra-and intercellular signaling (e.g., calcium signaling and signaling via reactive oxygen species), and the biosynthesis of signaling molecules (e.g., lipids, hormones, and neurotransmitter intermediates) (Devine and Kittler, 2018). Perhaps perturbed transport of mitochondria to the pre-synapse of peripheral neurons in CMT provides an opportunity to better understand other neurodegenerative disorders associated with defective presynaptic mitochondria, including Alzheimer's, ALS, Parkinson's, Friedreich's Ataxia, and Hereditary Spastic Paraplegia (Devine and Kittler, 2018).
But even considering the impaired mobility model for CMT within a single organism has some issues. It is difficult to imagine how reductions in mobility as high as 100% (Baloh et al., 2007;Rocha et al., 2018) could have a severe effect only on the longest axons, but not other axons that, while shorter than the peripheral neurons, are still very long compared to the ∼25 nm step size of a motor protein.
When considering these conundrums, it is helpful to consider alternative mechanisms for replenishing mitochondria in neurons, many of which are reviewed elsewhere (Misgeld and Schwarz, 2017;Yu and Pekkurnaz, 2018). Briefly, mitochondrial rejuvenation is speculated to be at least partially mediated via local translation in the axon. Interestingly, multiple CMT mutations affect local translation machinery (Table 1). More recent work showed that mitochondria serve as a stable compartment for mediating biogenesis by serving as an energy source for synaptic translation (Rangaraju et al., 2019). This raises a chicken vs. egg question: Does reduced mitochondrial mobility impair local translation needed for synaptic and therefore neuronal health and maintenance? Or does impaired local translation lead to dysfunctional mitochondria that cannot be replaced without sufficient mobility? That CMT is caused by mutations disrupting both mobility and local translation indicates these two processes have a unique relationship in long axons.

CONCLUSION AND OPEN QUESTIONS
One open question is how the overall distribution of mitochondria is altered in CMT neurons, and how this relates to axonal maintenance. A recent study showed mitochondria tend to distribute along the length of axons with regular spacing, and that inter-mitochondrial feedback mediates their positioning and movement (Matsumoto et al., 2020). Is this feedback-based positioning altered in CMT? Do mutations affecting mobility result in CMT via a "domino effect" caused by defects in relatively local repositioning between axonal mitochondria, which then cascades with increasing defects as a function of increasing axonal length? How much longer does it take mitochondria in CMT patients to traverse the entire length of an axon? Defective mitophagy has been implicated in other neurodegenerative disorders and some studies have linked CMT to alterations in autophagy (Colecchia et al., 2018;Gautam et al., 2019). Is there a reduction in the turnover rate of mitochondria in CMT patients? Mitochondria are increasingly being implicated as important players in adaptive and innate immune responses and inflammatory pathology, including neurodegeneration (West, 2017;Newman and Shadel, 2018). Could "mitoflammation" contribute to the pathophysiology of CMT? How do any and all of these factors affect mitochondria at the pre-synapse of CMT peripheral neurons, likely the most important subpopulation of mitochondria in these cells? These are surprisingly open questions we expect to be addressed in the coming years using animal and cell models of CMT.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

FUNDING
The Waitt Advanced Biophotonics Center was funded by the Waitt Foundation and Core Grant application National Cancer Institute (NCI) Cancer Center Support Grant (CCSG; CA014195). GS is the Audrey Geisel Chair in Biomedical Science and supported by NIH R01 AR069876. UM is a Chan Zuckerberg Initiative Imaging Scientists and supported by NIH R21 DC018237. CS was supported by NIH F32 GM137580.

ACKNOWLEDGMENTS
We would like to thank Timothy Shutt for critical feedback on the original version of the manuscript. We are grateful to Amy Cao (Salk Institute for Biological Studies) for help with the cartoon diagram in Figure 2.