Fig4 Deficiency: A Newly Emerged Lysosomal Storage Disorder?

Abstract

FIG4 (Sac3 in mammals) is a 5’-phosphoinositide phosphatase that coordinates the turnover of phosphatidylinositol-3,5-bisphosphate (PI(3,5)P₂), a very low abundance phosphoinositide. Deficiency of FIG4 severely affects the human and mouse nervous systems by causing two distinct forms of abnormal lysosomal storage. The first form occurs in spinal sensory neurons, where vacuolated endolysosomes accumulate in perinuclear regions. A second form occurs in cortical/spinal motor neurons and glia, in which enlarged endolysosomes become filled with electron dense materials in a manner indistinguishable from other lysosomal storage disorders. Humans with a deficiency of FIG4 (known as Charcot-Marie-Tooth disease type 4J or CMT4J) present with clinical and pathophysiological phenotypes indicative of spinal motor neuron degeneration and segmental demyelination. These findings reveal a signaling pathway involving FIG4 that appears to be important for lysosomal function. In this review, we discuss the biology of FIG4 and describe how the deficiency of FIG4 results in lysosomal phenotypes. We also discuss the implications of FIG4/PI(3,5)P₂ signaling in understanding other lysosomal storage diseases, neuropathies, and acquired demyelinating diseases.

Introduction

Phosphatidylinositol 3,5-bisphosphate (PI(3,5)P₂) is one of seven known phosphoinositides that are involved in the regulation of intracellular organelle trafficking in mammalian cells (Figure 1A). One of the key proteins that regulates the concentration of PI(3,5)P₂ is known as the FIG4 phosphatase (alternatively known as SAC3 in mammals). Recent discoveries of recessive FIG4/SAC3 gene mutations causal for neurodegeneration, abnormal lysosomal storage, and segmental demyelination in humans have substantiated its importance in neuronal functions and myelination. These pathological findings during FIG4 deficiency are reminiscent to what has been described in other lysosomal storage diseases, supporting the role of endolysosomal pathway dysregulation in the neurological disease. In this review, we will first discuss the clinical presentation, electrophysiological alterations, and endolysosomal pathology in patients and rodents with Fig4 deficiency. We will then describe how FIG4/SAC3 controls PI(3,5)P₂ abundance and endolysosomal trafficking in tandem with other PI(3,5)P₂ regulatory proteins. Finally, we will discuss how these molecular events might involve in the abnormal lysosomal accumulation in this neurological disease. For in-depth reviews of Fig4 and PI(3,5)P_2, refer to (Dove et al., 2009a; Ho et al., 2012).

Figure 1.

A. A diagram of phosphotidylinositide metabolism. PI = phosphotidylinositol. Relevant enzymes to this paper are marked. Their related diseases are shown in parenthess. Notice that all pathways involved in 4-phosphate are simplified since they are not relevant to this review. B. A hypothetical mechanism in Fig4 deficiency is illustrated.

Clinical Presentation of CMT4J

In humans, the FIG4/SAC3 gene is localized to chromosome 6q21 and it encodes a protein of approximately 103 kD comprised of 907 amino acids (Chow et al., 2007). CMT4J is an inherited peripheral neuropathy that is caused by autosomal recessive mutations of the FIG4 gene. Typically, the disease is caused by the combination of a FIG4 null allele and a missense FIG4^I41T mutation in the other allele (Chow et al., 2007; Nicholson et al., 2011; Zhang et al., 2008). These recessive mutations result in a loss of function of FIG4/SAC3 and a deficiency of PI3,5P₂. While the missense mutant FIG4^I41T allele still produces a partially functional phosphatase, the mutation results in an increase of FIG4/SAC3 cytosolic degradation (Ikonomov et al., 2010). Thus, the I41T point mutation consequently exerts a loss of function effect (Lenk et al., 2011).

The deficiency of FIG4/SAC3 in CMT4J patients manifests in rapidly progressive asymmetric muscle weakness with minimal or absent sensory symptoms. This renders CMT4J as a special subtype of CMT since most forms of CMT manifest as slowly progressive, symmetric and length-dependent polyneuropathies (Patzko and Shy, 2011). Symptomatic onset of CMT4J can range from early childhood to adulthood. Similarly, the severity of the CMT4J phenotype can vary from mild neurological impairment to fatal outcomes (Nicholson et al., 2011). It is currently unknown what genetic factors, if any, modify the onset and severity of CMT4J. While early onset patients can develop gait abnormalities with asymmetric limb involvement that appear as soon as the patients begin walking in childhood, late onset patients may develop subtle abnormalities for years that are insufficient to prompt them to seek medical attention (Nicholson et al., 2011; Zhang et al., 2008).

The most prominent clinical finding in CMT4J patients is typically muscle weakness with atrophy. In almost all cases, the severity of the weakness/atrophy is conspicuously asymmetric. There are mild sensory abnormalities, such as reduced sensation to touch and/or pin-prick in distal limbs. Deep tendon reflex may also be mildly decreased. Interestingly, however, central nervous system abnormalities have not been observed in patients with CMT4J (Nicholson et al., 2011; Zhang et al., 2008). This issue will be discussed further in the next section. On neurological examination, dysfunction of the cranial nerves is uncommon and these patients usually display normal cognitive function. Their brain MRI scans show no abnormalities (Nicholson et al., 2011; Zhang et al., 2008).

Most CMT4J patients have been examined by electrophysiological studies. Nerve conduction studies show prolonged distal latencies and reduced conduction velocities indicative of demyelination. Unlike most CMT1 patients, who have uniform slowing of conduction velocities with little variation from one nerve to another or from one limb to another, CMT4J patients have decreased conduction velocities that are often non-uniform or asymmetric. Temporal dispersion occurs in some cases. These findings closely resemble the electrophysiological features of acquired demyelinating diseases such as chronic inflammatory demyelinating polyneuropathy and Guillain-Barre syndrome.

This demyelinating change is further supported by analyses of human sural nerve biopsies from patients with CMT4J. Myelin thickness appears reduced in some myelinated nerve fibers and onion bulb is occasionally observed. There is also a severe loss of large myelinated nerve fibers (Nicholson et al., 2011; Zhang et al., 2008). In a previous study, we analyzed a sural nerve biopsy that was taken nine years prior to the onset of severe weakness and muscle denervation. A teased nerve fiber study on the sural nerve biopsy demonstrated segmental demyelination in 60% of the myelinated nerve fibers (Zhang et al., 2008). Thus, it is possible that patients with CMT4J may develop a primary demyelinating neuropathy or a neuropathy with a predominant pathology of demyelination in the early phase of its clinical course.

By the time that severe weakness and even paralysis develops in CMT4J patients, needle electromyography demonstrates abundant and diffuse abnormal spontaneous activities such as fibrillations and positive waves in muscles. Motor unit action potentials are complex in their morphology with reduced recruitment. These findings, along with the asymmetric progression and minimal sensory symptoms noted earlier, support a phenotype of spinal motor neuron degeneration or motor neuron disease in the later phase of CMT4J (Zhang et al., 2008). For a more complete evaluation of CMT4J pathology, autopsies of CMT4J patients may be necessary in order to understand the extent of spinal motor neuron degeneration and lysosomal storage pathology (discussed in the following sections).

The Pale Tremor Mouse

The characterization of a spontaneous mouse mutant of Fig4/Sac3 deficiency, the pale tremor (plt) mouse, has allowed scientists to gain a deeper insight into the neurodegenerative pathology observed in CMT4J patients. Plt mice are homozygous null for Fig4 due to a truncation mutation that is produced through transposon insertion and they display a 40–71% reduction of fibroblast PI(3,5)P₂ levels (Chow et al., 2007; Vaccari et al., 2011). Much like that observed in CMT4J patients, plt mice exhibit muscle weakness, severe loss of myelinated large diameter nerve fibers in the sciatic nerve and dorsal/ventral roots, loss of dorsal root ganglion sensory neurons, and segmental demyelination. Unlike patients with CMT4J, however, plt mice undergo extensive neuronal degeneration in cortical and subcortical regions (Chow et al., 2007; Zhang et al., 2008), and display reduced numbers of myelinating oligodendrocytes (Winters et al., 2011). When compared with CMT4J patients who can live well into adulthood, plt mice die relatively early in their lifespans at around 4–6 weeks. The absence of major CNS pathology in CMT4J patients is presumed to stem from the residual activity of the FIG4^I41T protein, which plt mice do not possess (Lenk et al., 2011; Zhang et al., 2008).

Additional variations of the plt mouse model have been recently developed, such as Fig4 null mice that express Fig4/Sac3 specifically in neurons. Using this model, it was shown that neuronspecific expression of Fig4/Sac3 is able to rescue CNS neurodegeneration and sciatic nerve demyelination (Ferguson et al., 2012). Such findings have established the plt mouse and its variations as valuable models for investigating how Fig4 affects neuronal survival and myelination.

The Lysosomal Pathology of Fig4 Deficiency: Implications for Lysosomal Storage Diseases

Fig4/Sac3 deficiency causes major pathological changes to lysosomes in the nervous system. There are two types of abnormal lysosomal storage in Fig4 deficient cells that likely stem from the dysregulation of endolysosomal trafficking, a process in which Fig4/Sac3 function is critically involved. The first occurs in Fig4/Sac3 deficient sensory neurons, which accumulate numerous “bubble-like” vacuoles, which are derived from dilated endolysosomes. These vacuoles accumulate in perinuclear regions and interfere with intracellular organelle trafficking processes (Katona et al., 2011; Zhang et al., 2008). The second occurs in Fig4 deficient cortical/spinal motor neurons and glia, in which dysfunctional lysosomes abnormally store large quantities of lipids and proteins (Katona et al., 2011). These cellular phenotypes are reminiscent to those of other lysosomal storage diseases such as Niemann Pick C, Tay-Sachs, and Mucolipidosis type IV. Thus, Fig4/Sac3 deficiency may become a new member of lysosomal storage disorders. Although it is still unclear as to how these pathological phenotypes contribute to neurodegeneration during Fig4/Sac3 deficiency, the study of Fig4/Sac3 deficient pathogenesis may potentially reveal disease mechanisms implicated in various lysosomal storage disorders. In the following sections, we will summarize the pathological data regarding endolysosomal defects in the Fig4-deficient plt mouse model and also examine the implications of these pathological changes in diseases similar to CMT4J.

a. Abnormal accumulation of vacuolated endolysosomes in sensory neurons

Accumulation of vacuoles occurs in 42–82% of plt dorsal root ganglion (DRG) sensory neurons in vitro and in vivo as early as postnatal day 4 (Figure 2). This observation is consistent with the fact that Fig4 expression is persistently high in wild-type DRG sensory neurons throughout development and adulthood, but is only transiently high in spinal motor neurons during early development. By adulthood, Fig4/Sac3 expression is hardly detectable in wild-type spinal motor neurons (Guo et al., 2012). One might wonder how severe vacuolization can occur predominantly in sensory neurons, yet CMT4J patients typically do not complain about sensory symptoms. A possible explanation may be that there is an inhomogeneous degeneration of sensory neurons that preserves non-myelinated smaller nerve fibers responsible for pain and pinprick sensations.

Figure 2.

Vacuoles in plt neurons of the DRG. (A) A semithin section was processed from the DRG of a 5-week-old wild-type mouse. The DRG is filled with sensory neurons with a polygon shape and pale nucleus. There are a few myelinated nerve fibers visible between the neurons. (B) The same study was done in the DRG of a 5-week-old plt mouse. Many vacuoles are visible in the neurons (arrowhead). The cytoplasm in some neurons is fully occupied by these vacuoles. (C) Two adjacent vacuoles were examined under electron microscopy, and were determined to be single-membrane bound. Vacuoles have a watery appearance. They contain oil-drop-like materials. A few pieces of membranous debris may occasionally be seen in the vacuoles. (D) A semithin section of DRG from a wild-type mouse at P4 was imaged under light microscopy (scale bar = 20 µm). (E) Large vacuoles in plt DRG neurons are readily identifiable at P4 (arrows). Note: Reprinted with permission from “Distinct pathogenic processes between Fig4- deficient motor and sensory neurons,” by Katona et al., 2011, European Journal of Neuroscience, 33, pp. 1401-1410.

The abnormal storage of cytoplasmic vacuoles in Fig4/Sac3 deficient sensory neurons has been suggested to occur as a result of dysfunctional lysosomal trafficking processes. In Fig4-deficient sensory neurons, the vacuoles become clustered in perinuclear regions and block intracellular organelle trafficking (Zhang et al., 2008). These vacuoles arise from the late endosomes and lysosomes of plt mice and CMT4J patients, as revealed by immunohistochemistry in which vacuoles co-localize with a late endosome/lysosome marker, but not an early endosome marker (Figure 2) (Chow et al., 2007; Zhang et al., 2008). Such localization is interesting, given that Fig4/PI3,5P₂ are thought to aid in the maturation of early endosomes to late endosomes.

b. Abnormal endolysosomal storage of high electron dense materials in spinal motor neurons and glia

The lysosome is a critical cellular compartment that mediates the degradation of cellular debris and other materials from the phagocytic, endocytic, and autophagic pathways (Luzio et al., 2007). Failure of lysosomes to degrade and/or properly recycle their contents has been increasingly implicated in a variety of neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, Huntington’s disease, and amyotrophic lateral sclerosis (Nixon et al., 2008; Zhang et al., 2009). Ultrastructural studies of the plt brain and spinal cord have documented an excessive accumulation of large endolysosomes containing electron-dense granules in up to 80% of spinal motor/cortical neurons and glia (Figure 3). These granules localize to perinuclear regions in a manner similar to the vacuolated endolysosomes of Fig4-deficient sensory neurons. Other neurodegenerative lysosomal storage disorders, such as lipofuscinosis and Tay-Sachs disease, also possess perinuclear vacuole accumulation (Katona et al., 2011; Myerowitz et al., 2002; Sulzer et al., 2008). Additionally, the abnormal accumulation of lysosomal proteins LAMP1, LAMP-2, and NPC1 has been documented in the brains of plt mice (Ferguson et al., 2009; Ferguson et al., 2012; Katona et al., 2011). In wild-type samples, only 40% of neurons and glia possess granules, and these granules are less abundant, smaller in size, less electron-dense, and more dispersed throughout the cytoplasm than those of plt mice (Katona et al., 2011). This accumulation of high-electron dense materials has also been documented in plt spleen cells as well, consistent with a lysosomal storage disorder. These electron-dense granules are thought to accumulate in dysfunctional lysosomes that lack lumenal acidification and degradative abilities (Katona et al., 2011). Thus, there is a lack of vacuolar accumulation in spinal motor neurons until the end stages of disease progression (~postnatal week 5 in plt mice) (Katona et al., 2011). Even at later stages, vacuolar accumulation in spinal motor neurons is comparatively less than that observed in the DRG sensory neurons (Chow et al., 2007; Katona et al., 2011). The potential cause for the two different forms of lysosomal storage will be discussed further later.

Figure 3.

Mouse spinal cord and brain were dissected for electron microscopy. (A) A motor neuron was identified from the spinal anterior horn of a wild-type mouse. The sizes of many motor neurons are larger than those in the plt anterior horns. This motor neuron contains electron-dense granules (arrow). However, these granules are small and scattered throughout the cytoplasm, rather than clustered in the perinuclear region. (B) In contrast, these high electron dense granules are found in many plt neurons and/or glia. They are numerous and always clustered in the perinuclear region (arrow). Granules are not detectable in axons (asterisk). (C) Granules were inspected under high magnification. They are inhomogeneous. Large granules appear to be fused from several dark vesicles. (D) Similar accumulations were observed in cerebral neurons. Note: Reprinted with permission from “Distinct pathogenic processes between Fig4-deficient motor and sensory neurons,” by Katona et al., 2011, European Journal of Neuroscience, 33, pp. 1401-1410.

c. The lysosomal storage pathology of neurons and glia is likely pathogenic to neurodegeneration during Fig4 deficiency

Lysosomal diseases were initially defined as a group of diseases with inborn errors of metabolism that displayed pathological inclusions (abnormal lysosomal storage) formed by intravacuolar storage of heterogeneous materials in cells (Hers, 1965). The cause for these diseases was originally thought to be a result of a deficiency in certain enzymatic activity. Examples of this type would include mucopolysaccharidoses (such as Hurler-I or Sanfilippo-III disease), sphingolipidoses (such as Tay-Sachs, Gaucher or Krabe’s disease) and glycoproteinoses (α-Fucosidosis or α-Mannosidosis). With the advance of molecular genetics and cell biology, this view has proven to be over-simplistic. Defective enzyme activity has also been found to be caused by genetic factors that regulate the synthesis or transport of lysosome-related proteins (Schultz et al., 2011). Thus, these defects emerge as a fourth group (miscellaneous group) of lysosomal diseases in addition to the three aforementioned groups of lysosomal diseases. For instance, loss of LAMP-2 function in patients with Danon disease develops a lysosomal disorder that predominantly affects cardiac and skeletal muscles with profound accumulation of autophagosomes in the muscle cells (Saftig et al., 2010). LAMP-2 is not a lysosomal enzyme, but rather a lysosomal membrane protein that appears critical for certain substrates to be delivered into lysosomes via a chaperone-mediated autophagic pathway (Orenstein and Cuervo, 2010; Saftig et al., 2010). We believe that it is reasonable to suggest that Fig4 deficiency belongs in this fourth group of lysosomal storage disorders due to the following reasons.

First, our study has provided pathological evidence showing an accumulation of high electron dense organelles containing heterogeneous materials in Fig4/Sac3 deficient cells, a morphological finding consistent with abnormal lysosomal storage. This lysosomal accumulation is also accompanied by high levels of lysosomal protein LAMP-2 (Katona et al., 2011). While such accumulation has not been shown in human neurons, however, vacuoles derived from dilated endolysosomes have been demonstrated to accumulate in fibroblasts isolated from humans with CMT4J (Zhang et al., 2008). It is important to be reminded that vacuolated lysosomal storage has been observed in other lysosomal storage diseases, such as Mucolipidosis, and can be interchangeable with high electron dense lysosomal storage under certain conditions (Kogot-Levin et al., 2009). Nevertheless, it would be helpful if the electron dense form of lysosomal storage could also be demonstrated in human neurons or glia when autopsy materials become available. Second, there are multifaceted features in Fig4−/− mice that are consistent with lysosomal disease, including diluted coat color, high electron dense lysosomal storage in the spleen, growth retardation, and hydrocephalus. Third, patients with CMT4J mainly show abnormalities in the peripheral nerves and spinal cord, but their CNS is spared. This should not be viewed as evidence against lysosomal disease since other classical lysosomal diseases can be confined exclusively to the peripheral nerves and/or spinal cord. For instance, Adrenomyeloneuropathy (AMN) is a well-known variant of Adrenoleukodystrophy that mainly affects spinal cord and peripheral nerves. Clinical cerebral abnormalities may develop in patients with AMN, but are later features and only occur in 45% of patients with AMN (Edwin et al., 1990; Moser, 1997). Moreover, an isolated peripheral neuropathy has also been observed in another lysosomal disease known as Krabbe’s disease (Marks et al., 1997). Finally, PI3,5P₂ is decreased in Fig4 deficient cells (Chow et al., 2007; Vaccari et al., 2011). This phosphoinositide is well documented to play a significant role in regulating the trafficking of lysosomal membrane. Loss of function of proteins in complex with Fig4/Sac3, such as Fab1/PIKfyve and Vac14/ArPIKfyve, all result in abnormal lysosomes (Di Paolo and De Camilli, 2006; Ho et al., 2012; Zhang et al., 2007). While vacuolated lysosomes have been reported in Fab1/PIKfyve or VAC14/ArPIKfyve deficient rodents (Ikonomov et al., 2011; Zhang et al., 2007), it would be interesting to examine whether the abnormal storage of high electron dense lysosomes is also present in the neurons of these animals.

Abnormal lysosomal storage in Fig4 deficient motor neurons and glia appears to be closely associated with neurodegeneration. Although Sac3/Fig4 deficiency causes lysosomal defects in both sensory and motor neurons/glia (Katona et al., 2011; Zhang et al., 2008), the clinical manifestations appear to correlate with the degeneration of spinal motor neurons and demyelination of the peripheral nerves. This suggests that abnormal lysosomal storage of protein and lipids in spinal motor neurons and glia is the major factor responsible for the CMT4J clinical phenotype. Such a pathogenic mechanism is not particularly uncommon, since spinal motor neuron degeneration and de-/dysmyelination of peripheral nerves have also been observed in a number of lysosomal storage disorders. For example, in a group of lysosomal storage disorders known as the GM2 gangliosidoses, patients may have a deficiency of a lysosomal hydrolase (β-hexosaminidase). These patients present with clinical features of spinal muscular atrophy that primarily result from the degeneration of spinal motor neurons (Navon et al., 1997). Furthermore, the knockout of β-hexosaminidase in mice also results in the abnormal lysosomal storage of glycosphingolipid that accompanies neuronal degeneration, including damage to spinal motor neurons (Liu et al., 1999). Removal of lysosomal glycosphingolipid accumulation significantly improves life span, neurological function, and reduces the degeneration of neurons (Liu et al., 1999). These findings suggest that abnormal lysosomal storage can be causal for neuronal degeneration, and that it could be the primary pathogenic mechanism resulting in the degeneration of Fig4 deficient spinal motor neurons.

How does the Fig4 deficiency lead to abnormal lysosomal storage (either electron lucent or dense form)?

The prominent lysosomal storage in Fig4/SAC3 deficiency raises an important question: how does the deficiency of Fig4/SAC4 result in lysosomal storage? Based on current literature, we describe several potential pathways below.

Fig4/SAC3 is a SAC-domain phosphoinositide phosphatase with specificity toward the 5’- phosphate of PI(3,5)P₂. Structural studies of Fig4/SAC3 have not yet been performed, but have been inferred through its homologous protein in S. cerevisiae known as Sac1p. Sac1p has a catalytic domain containing the conserved catalytic motif CX₅R(T/S) specific for phosphoinositide substrates. The catalytic core shares a similar topology with other phosphoinositide phosphatases like MTMR2 and PTEN (Table 1), but possesses a unique configuration in its P-loop domain which contains the catalytic CX₅R(T/S) motif (Manford et al., 2010).

Table 1.

Catalytic enzymes for phosphatidylinositide 5-phosphate.

Protein	Substrate	Knockout Mouse	Human Disease
PAS complex PIKfyve	PI, PI3P	embryonic lethal at stage of 32–64 cells; vacuoles in cells (Ikonomov et al JBC 2011)	Francois-Weatens-Fleck corneal dystrophy (CFD) by heterozygous loss of function of PIKfyve – white fleck in all layers of stroma(Li et al Am J Hum Genet 2005)
ArPIKfyve		death at P1-2; vacuoles in brain, DRG and fibroblasts (Zhang et al PNAS 2007)	none
SAC3/FIG4	PI3,5P2	death at P5-6 weeks; vacuoles in DRG and fibroblasts; electron dense lysosomal storage in neuron, glia; segmental demyelination (Chow et al Nature 2007;Katona et al Euro J Neurosci 2011)	CMT-4J by homozygous loss of function of FIG4 – progressive asymmetric weakness, minimal sensory symptoms, segmental demyelination in PNS (Zhang et al Brain 2008;Nicholson et al Brain 2012)
Phosphatidylinositide 5-phosphate phosphatases
5-phosphatase-1	I1,4,5P3; I1,3,4,5P4	only for soluble inositol polyphosphates – no further discussion
5-phosphatase-II /INPP5b	PI4,5P2 PI3,4,5P3; IP3; IP4	male sterility, vacuoles in seminiferous tubule epithelium (Hellstein et al Biol Reprod 2002)	none
OCRL	same as INPP5b	no phenotype (Janne et al JCI 1998)	Lowe syndrome - congenital cataracts, neonatal hypotonia, behavior disorder, renal tubule dysfunction
Synaptojanin-1	PI3P, PI4P, PI5P	weakness, ataxia, convulsion vesicle accumulation in nerve terminals (Cremona et al Cell 1999)	synpatojanin-1 increased in Down’s syndrome (Arai et al Brain Dev 2002)
Synaptojanin-2	PI3,5P2	none	none
SHIP-1	PI1,3,4,5P4; PI3,4,5P3; PI1,2,3,4,5P5	none	none
SHIP-2/INPPL1	PI3,5P2; IP3; IP4	resistant to dietary obesity (Sleeman et al, Nat Med 2005)	genetically linked to patients with diabetes or insulin resistance
SKIP	PI4,5P2 PI3,4,5P3	embryo lethal in skip−/− mouse; increased glucose tolerance & insulin sensitivity (Ljuin et al Mol Cell Biol 2008)	none
PIPP	PI3,4,5P3 PI4,5P2 PI1,4,5P3	none	none

In order for Fig4/Sac3 to exert its biological function, it typically complexes with a scaffolding protein known as Vac14/ArPIKfyve and a 5’-kinase of PI3P known as Fab1/PIKfyve (Duex et al., 2006; Efe et al., 2005; Efe et al., 2007; Ikonomov et al., 2009; Jin et al., 2008). This regulatory complex (PAS) is thought to localize on early endosomal membranes during the transition from early endosomes to late endosomes (Guo et al., 2012; Sbrissa et al., 2007; Yuan et al., 2007). Alongside the myriad of endosomal maturation processes that occur at this time, the PAS complex is suggested to mediate the conversion of early endosomal PI3P to late endosomal PI(3,5)P₂ - a conversion that appears to be essential for protein sorting and the trafficking of late endosomes to the degradative lysosomal compartment (Huotari and Helenius, 2011; Sbrissa et al., 2007). The I41T mutation that underlies CMT4J is located in the SacN domain, which is suggested by the structural study to interact with other proteins (Manford et al., 2010). Interestingly, this mutation has been shown to affect the interaction between Fig4 and Vac14/ArPIKfyve, promoting FIG4^I41T degradation (Ikonomov et al., 2010).

Although the abundance of PI(3,5)P₂ is much lower in comparison to other phosphoinositides (PI3,5P₂ comprises just 0.04–0.08% of total cellular inositol phospholipid), cells are very sensitive to PI(3,5)P₂ loss. Once PI(3,5)P₂ has formed on late endolysosomal membranes, it interacts with a variety of proteins that may mediate endolysosomal acidification (Dove et al., 2009b; Gary et al., 1998; Rusten et al., 2006), the hyperosmotic stress response (Bonangelino et al., 2002; Dove et al., 1997; Sbrissa and Shisheva, 2005), endolysosomal membrane fusion and fission (Gary et al., 1998; Ikonomov et al., 2006; Ikonomov et al., 2009; Poccia and Larijani, 2009), ryanodine receptor activation (Touchberry et al., 2010), Ca²⁺ regulation (Dong et al., 2010; Shen et al., 2012; Shen et al., 2009; Silswal et al., 2011), exocytosis (Berwick et al., 2004; Dove et al., 2009b), gene expression (Han and Emr, 2011), and trafficking between the endolysosome and Golgi apparatus (Bryant et al., 1998; Dove et al., 2009b; Rutherford et al., 2006). Each component of the Fig4/Vac14/Fab1 complex is dependent on either direct or indirect interactions within the complex for proper function and stability. For example, it has been shown that Fig4 must associate with Vac14/ArPIKfyve before the complex can efficiently bind Fab1/PIKfyve (Botelho et al., 2008; Ikonomov et al., 2009; Sbrissa et al., 2007). In this manner, Fig4/Sac3 can decrease PI(3,5)P₂ levels via its phosphatase function and also promote PI3,5P₂ synthesis by acting as a secondary scaffold for the Fab1/Vac14 interaction. However, the later function appears dominant. Loss of either component results in a similar outcome – a destabilization of the PAS complex and a reduction of PI(3,5)P₂ that gives rise to greatly dilated endolysosomes (Chow et al., 2007; Gary et al., 1998; Ikonomov et al., 2001; Jin et al., 2008).

Our studies have localized the Fig4/SAC3 protein to early endosomal membranes, where it may form the PI(3,5)P₂ regulatory complex with Vac14 and Fab1 in order to regulate early endosome-to-lysosome maturation (Guo et al., 2012). It has also been suggested that PI3P and PI(3,5)P₂ on early endosomal membranes are involved in the recruitment of ESCRT protein machinery that regulates the biogenesis of multi-vesicular bodies (intermediate organelles between early endosomes and late endosomes) (Bonangelino et al., 1997; Ho et al., 2012; Ikonomov et al., 2002b). Given these findings, one might speculate that the deficiency of Fig4/SAC3 and PI(3,5)P₂ could impair endolysosomal membrane fusion/scission and protein sorting processes required to properly target proteins and lipids for lysosomal degradation. Based on studies in yeast, lysosomal fission is more likely affected by Fig4/PI(3,5)P₂ deficiency than lysosomal fusion with other organelles (Dove et al., 2009a). This notion is also supported by our previous finding that showed a decrease of organelle fission once the dilated endolysosomes were fused into the clusters of perinuclearly accumulated vacuoles in human fibroblast cells of CMT4J (Zhang et al., 2008).

One caveat in these proposed mechanisms is that they are all built on an unproven premise – which is that the reduction of PI(3,5)P₂ is directly responsible for the storage of the dilated endolysosomes. In PIKfyve deficient culture cells, endolysosomal vacuoles can be eliminated by injecting PI(3,5)P₂ into these cells, suggesting that PI(3,5)P₂ deficiency is causal for the vacuolation (Ikonomov et al., 2002a). However, this causal role of PI(3,5)P₂ in Fig4 deficient cells has yet to be established. Moreover, fibroblasts from plt mice were recently reported to have ~40% less PI(3,5)P₂ than wild-type (Vaccari et al., 2011), which is in contrast to a previous finding of a 71% decrease in PI(3,5)P₂ reported by Chow et al (Chow et al., 2007). This may be problematic, since a ~40% decrease in PI(3,5)P₂ was found in fibroblasts from mice with a single PIKfyve allele knockout that did not show a phenotype (Ikonomov et al., 2011). Such a discrepancy could be a result of technical difficulties in the measurement of PI(3,5)P₂ levels due to its extremely low abundance. Alternatively, there could be a PI(3,5)P₂-independent molecular mechanism causing vacuolated endolysosomal accumulation. It would be important to compare the anatomical localizations of Fig4 and other proteins in the PAS complex (PIKfyve and ArPIKfyve) along with the pathological changes between plt mice and animal models with loss of function of other proteins in the PAS complex. Identical localization and pathological changes between these proteins would support a shared common mechanism, such as PI(3,5)P₂ deficiency.

The recent discovery of a lysosomal ion channel activated by PI(3,5)P₂ provides an alternative pathogenic route for the abnormal lysosomal storage of Fig4 deficiency. This channel, known as the transient receptor potential mucolipin-1 (TRPML1/MCOLN1) channel, releases Ca²⁺ from the lysosomal lumen into the cytosol when PI(3,5)P₂ binds to its N-terminus (Dong et al., 2010). Mutation or inhibition of this channel results in decreased lysosomal Ca²⁺ release in vitro. Mutations in the gene encoding TRPML1/MCOLN1 in humans cause an autosomal recessive lysosomal storage disease; Mucolipidosis type IV (Bargal et al., 2000; Shen et al., 2012). In addition, down-regulation of TRPML1/MCOLN1 activity appears to underlie lysosomal defects in the Niemann Pick C lysosomal storage disorder (Shen et al., 2012). Similar to that observed in CMT4J patients and plt mice, cells that lack TRPML1/MCOLN1 exhibit enlarged endolysosomes and LEL-to-Golgi retrograde trafficking defects (Dong et al., 2010; Shen et al., 2012). Given the role of Fig4/PI3,5P₂ in endolysosomal regulation, one might suggest that the loss of Fig4 could subsequently lead to the inhibition of the TRPML1/MCOLN1 channel. In support of this view, transfection of TRPML1/MCOLN1 into Vac14/ArPIKfyve null fibroblasts with PI3,5P₂ deficiency can rescue the vacuolization phenotype (Shen et al., 2012). One might question this mechanism since a different type of lysosomal storage with membranous inclusions (in contrast to electron dense or translucent inclusions in Fig4−/− cells) has been observed in cells, mice, and patients with TRPML1/MCOLN1 deficiency (Folkerth et al., 1995; Miedel et al., 2008; Venugopal et al., 2007). However, Niemann Pick-C develops polymorphic lysosomal storage (membranous inclusions + electron dense + translucent vacuoles). Yet, a decrease of TRPML1/MCOLN1 activity appears responsible for the lysosomal defects in the Niemann Pick- C cell model (Shen et al., 2012). Interestingly, TRPML1/MCOLN1 has been suggested to play a role in neuronal autophagy (Curcio-Morelli et al., 2010), whereas, as to be discussed below, autophagy is not essential for the pathogenesis of CMT4J. Moreover, the clinical phenotype in patients with Mucolipidosis type IV differs from those in patients with CMT4J in multiple aspects (Nicholson et al., 2011; Wakabayashi et al., 2011; Zhang et al., 2008). Thus, it is likely that TRPML1/MCOLN1 is involved in additional functions and/or could also be expressed in different tissues/cellular localizations when compared with Fig4. It would be helpful to determine whether TRPML1/MCOLN1 is down-regulated in Fig4−/− cells and if over-expression of TRPML1/MCOLN1 can rescue the phenotype in Fig4 deficient cells.

Segmental Demyelination during Fig4 Deficiency

A nerve fiber can increase its action potential conduction velocity by wrapping its axon with Schwann cell membrane (myelin) in sequential segments (internodes). The segments are separated by punctuate gaps where the axon is denuded of myelin, called nodes of Ranvier. This configuration permits the action potentials to be conducted in an efficient saltatory fashion from one node to the next (Baker, 2002). Central to the myelination process are glial cells known as Schwann cells (located in the PNS) and oligodendrocytes (located in the CNS). These cells interact with axons and wrap them in electrically insulating myelin sheathing. Demyelination denudes axons and shunts depolarizing current out of the nerve. As a result, nerve conduction of action potentials can be slowed or even eliminated (The latter is also known as conduction block). This change causes neurological disabilities such as sensory loss, vision loss, and limb paralysis in a variety of neurological diseases (Relevant examples include Guillain-Barré Syndrome, chronic inflammatory demyelinating polyneuropathy, multiple sclerosis, certain subtypes of CMT, etc.) (Kaji, 2003).

Fig4 is strongly expressed in both oligodendrocytes and Schwann cells throughout development and adulthood, indicating that FIG4 may play an important role in myelination (Guo et al., 2012). Indeed, the deficiency of FIG4 results in segmental demyelination. Nerve conduction studies of plt sciatic nerves and CMT4J nerves demonstrated significantly reduced conduction velocities when compared to wild-type controls (Chow et al., 2007; Zhang et al., 2008). Oftentimes, these changes in conduction velocities were non-uniform. Additionally, segmental demyelination has been morphologically confirmed in >60% of teased CMT4J sural nerve fibers of an adult human and 11% of teased plt sciatic nerve fibers at the age of 5–6 weeks (Figure 4). Macrophage infiltration in the peripheral nerves also occurs in plt mice (Zhang et al., 2008). When combined with the knowledge that CMT4J manifests as an asymmetric and rapidly progressive weakness, these findings suggest that demyelination in CMT4J shares key features with acquired demyelinating diseases. Understanding how Fig4 deficiency impacts myelination could provide pathogenic insights applicable to other demyelinating diseases.

Figure 4.

Segmental demyelination in teased sciatic nerve fibres from plt mice. Mouse sciatic nerves were teased into individual nerve fibers and stained via immunohistochemistry. (A) Myelin specific protein MAG is strongly expressed in Schmidt-Lanterman incisures. MAG staining is absent in some segments of myelinated plt nerve fibers (e.g. between arrowheads in A,B). Labeling for NFp (phosphorylated neurofilaments) was intact in the same segment (C), demonstrating segmental demyelination. Many nuclei are lined up along this segment of the axon (D) that are derived from invading macrophages or remyelinating Schwann cells. Demyelination was not observed in wild-type nerve fibers. Note. Reprinted with permission from “Mutation of FIG4 causes a rapidly progressive, asymmetric neuronal degeneration,” by Zhang et al., 2008, Brain, 131, pp.1990-2001.

Reductions in myelinating oligodendrocyte (OL) abundance have recently been observed in the brain and optic nerve of plt mice. Additionally, this reduction in OL abundance is accompanied by a decrease in myelin thickness (Winters et al., 2011). These findings implicate Fig4 in the maturation of oligodendrocyte precursor cells (OPCs) to OLs. Exogenous Fig4 in neurons is sufficient to correct this phenotype, suggesting that dysmyelination in the CNS may be secondary to abnormal signaling from Fig4 deficient neurons. Although sciatic nerve demyelination appears to be rescued during neuron-specific Fig4 expression (Ferguson et al., 2012), it is still not known whether such a pathogenic mechanism applies to segmental demyelination in Fig4 deficient peripheral nerves. One might wonder how this mechanism could apply to CMT4J patients, given that they possess a Fig4^I41T allele that protects their CNS, but not their PNS, from neurodegeneration. Our recent studies of Fig4 expression during the development of wild-type mice indicate that Fig4 expression remains high in the PNS but drastically declines in the CNS with age (Guo et al., 2012). Thus, it is possible that PNS myelin may depend more on Fig4 activity than the myelin of the CNS.

Myelination in both the CNS and PNS appears to be sensitive to the levels of Fig4, displaying abnormalities even during Fig4^+/− heterozygosity. Recent studies of young Fig4^+/− mice have demonstrated transient reductions in CNS myelin thickness that are visible at P21 but absent by P90 (Winters et al., 2011). Our study found transiently reduced conduction velocities in the PNS of young Fig4^+/− mice which also became normal with age (Yan et al., 2012). Given that FIG4 heterozygosity has been documented in several patients with sporadic ALS, one might ask if the myelination abnormalities of heterozygous Fig4^+/− mice are relevant to ALS pathology (Chow et al., 2009). However, this speculation requires much evidence to substantiate since myelinspecific abnormalities have not been shown to be causal for ALS.

Given the endolysosomal abnormalities in Fig4 deficiency, it is reasonable to speculate that lysosomal dysfunction is involved in the myelination defects of CMT4J patients and plt mice. It should be emphasized that myelin abnormalities are a common feature in many lysosomal diseases (Folkerth et al., 1995; Marks et al., 1997; Moser, 1997). Another shared feature of many lysosomal storage diseases is an imbalance of intracellular calcium (Kiselyov et al., 2010). Excess intracellular calcium is known to cause demyelination (Smith et al., 1985; Smith et al., 2001; Smith and Hall, 1988). Considering that PI3,5P₂ is known to stimulate the lysosomal TRPML1/MCOLN1 channel to release calcium into the cytosol (Dong et al., 2010), and that PIKfyve is known to regulate Ca_v1.2 ion channel internalization (Tsuruta et al., 2009), a deficiency of Fig4 could lead to intracellular calcium abnormalities and subsequent demyelination.

In many lysosomal diseases, lysosomal storage in neurons is striking morphologically. In contrast, this storage is often much milder or even absent in myelin. This difference is also conspicuous in CMT4J. Given the high plasticity of Schwann cells (repetitive differentiation or dedifferentiation); one might wonder if accumulated lysosomal materials are efficiently discarded during demyelination.

Interestingly, an increase in PI(3,5)P₂ level is linked to autosomal recessive disorders such as CMT4B1 and CMT4B2, all of which display de-/dysmyelination. However, the increase of PI(3,5)P₂ level is associated with a very different type of myelin abnormality when compared with that in CMT4J. There are numerous enlarged paranodes in CMT4B1/2 myelinated nerve fibers. They are formed by excessive myelin folding primarily in the longitudinal direction of axons (Bolino et al., 2000; Bolino et al., 2004; Previtali et al., 2003; Quattrone et al., 1996; Vaccari et al., 2011). This myelin alteration is remarkably different from the typical segmental demyelination in Fig4 deficient myelinated nerve fibers. Together, these findings suggest that levels of PI(3,5)P₂ may have to be regulated within a tight range in order to maintain normal myelination.

Autophagy does not play a primary pathogenic role in Fig4 deficiency

Autophagy is essential for maintaining cell viability as it encompasses the lysosomal degradation of damaged organelles and the elimination of large protein aggregates (Ferguson et al., 2009; Klionsky, 2007). Progressive accumulation of protein aggregates has been found in numerous neurodegenerative diseases and can eventually become cytotoxic (Ross and Poirier, 2004). Mutations of Fig4 have been shown to result in the accumulation of autophagic intermediates (p62, LC3-II) in the astrocytes of plt mice. This accumulation is suggested to occur as a result of PI(3,5)P₂ deficiency that would normally coordinate the fusion of autophagosomal compartments (autophagosome to late endosome) and/or recycling of lysosomal membranes from mature autolysosomes (Ferguson et al., 2009). These findings raise the question of whether the loss of function of Fig4 impairs autophagy, thus playing a pathogenic role in CMT4J neurodegeneration.

This initial impression has since been challenged by several recent observations suggesting that autophagy is not a primary pathogenic mechanism in Fig4 deficiency. First, our studies of Fig4 expression throughout development found no expression of Fig4 in the astrocytes of wild-type rodents. Even after nerve injury which induced Fig4 up-regulation in many cells of the neural tissues, Fig4 expression was still absent in astrocytes (Guo et al., 2012). These results suggest that the accumulation autophagic intermediates in plt astrocytes is a secondary result occurring downstream of another pathogenic mechanism, such as astrocytic reaction to nearby neuronal degeneration. Moreover, we found that autophagosomes were either absent or only rarely detectable in Fig4−/− brain and or spinal cord by EM studies (Katona et al., 2011).

Second, a recent study evaluated the impact of Fig4 on the phenotype of mice null for Mtmr2 (a 3’-phosphoinositide phosphatase of PI3P and PI(3,5)P₂). Deficiency of Mtmr2 increases PI3,5P₂, however, no changes in the level of p62 or LC3-II/I ratio were observed in Mtmr2-null mice. Conversely, the double-knockout of Fig4 and Mtmr2 decreases PI3,5P₂ levels more than those in Fig4−/− tissues, but does not increase p62 levels relative to plt mice (Vaccari et al., 2011). These findings question the effect of PI3,5P₂ fluctuations on autophagic intermediate accumulation.

Third, additional research has shown that mice null for the PI3P kinase PIK3C3 display sensory neuron vacuolation with no observable impairment of neuronal autophagy. PIK3C3 catalyzes the production of PI(3)P and the loss of PIK3C3 would ultimately deplete PI3,5P₂. However, autophagy is not impaired in mice with PIK3C3 deficiency. Further knockout of a molecule necessary for canonical autophagy, Atg7, does not recapitulate the pathology of plt neurons (Chow et al., 2007; Katona et al., 2011; Zhou et al., 2010). Interestingly, while the Fig4/PI3,5P₂ pathway is highly conserved among different species, yeast does not require PI3,5P₂ for autophagy (Dove et al., 2009a; Gary et al., 1998).

Finally, a recent publication by Ferguson et al., 2012 further clarifies the autophagy issue. The authors utilized a global Fig4 null mouse with neuron-specific expression of Fig4 and demonstrated that neuron-specific expression of Fig4 is sufficient to rescue neuronal degeneration. In contrast, astrocyte-specific Fig4 expression did not rescue the phenotype, but restored the levels of autophagic intermediates. Although teased nerve fiber studies were not performed in this study, it also appears that demyelination in mouse sciatic nerves is rescued during neuron-specific Fig4 expression since slowed conduction velocities were reversed and gratios became comparable with wild-type mice. Together, these data strongly suggest that abnormal autophagy does not play a significant role in the pathogenesis of Fig4 deficiency. The accumulation of autophagic intermediates is mild and likely a secondary phenomenon.

Diverse functions of PI3,5P₂ kinases and phosphatases

Despite the miniscule amount of PI(3,5)P₂ in cells, catalytic enzymes involved the regulation of phosphatidylinositol 5’-phosphate can form an impressively long list. This list includes the three proteins in the PAS complex (Fab1/PIKfyve, Fig4/SAC3 and Vac14/ArPIKfyve) and ten phosphatidylinositol 5-phosphate phosphatases (Table 1). Thus, it is important to consider these enzymes while investigating the FIG4 functions, since there are potential functional overlaps between FIG4 and these 5-phosphate phosphatases. On the other hand, there have been multiple diseases identified which are caused by the loss of function of these enzymes. The phenotypes of these diseases are highly diverse. Moreover, knockout mice of different enzymes in the list have also revealed a variety of phenotypes in vivo. Surprisingly, many of these knockouts result in a lethal phenotype without being compensated by other enzymes that can perform a similar catalytic activity (Table 1). These findings not only demonstrate diverse functions of these enzymes in biological systems, but also suggest that some of these functions might be PI(3,5)P₂ independent. Alternatively, they share the same preference of certain substrates, but express in different tissues or cells to exert different biological functions. For instance, loss of function of OCRL by autosomal recessive mutations results in the X-linked oculocerebrorenal syndrome of Lowe. Patients with Lowe’s syndrome suffer primarily from congenital cataracts, neonatal hypotonia, joint hyper-extensibility, intellectual disability (self-abuse, agitation) and renal proximal tubule dysfunction (Loi, 2006). OCRL has catalytic preference toward PI(4,5)P₂ over PI(3,5)P₂. OCRL deficiency raises the level of PI(4,5)P₂ in cells. OCRL expresses ubiquitously in a variety of tissues and is localized to the Golgi/endosome system that utilizes the clathrin-coated vesicles for bidirectional trafficking between the two organelles. Interestingly, pathological studies have demonstrated single membrane bound, electron lucent vacuoles in OCRL deficient cells. Electron dense or membranous inclusions are only occasionally seen (Wisniewski et al., 1984). However, these inclusions are mainly derived from endosomes (Pirruccello and De, 2012).

A close family member of OCRL, INPP5E, expresses highly in brain and kidney, but within a cell, INPP5E has a localization and catalytic activity similar to that of OCRL. Recessive mutations of INPP5E have been found in a subset of patients with MORM syndrome, a ciliopathy affecting multiple organ systems that commonly causes mental impairment, truncal obesity, retinal dystrophy and micropenia (Jacoby et al., 2009). This phenotype is conspicuously different from that in Lowe’s syndrome. Moreover, synaptojanin-1 also shows catalytic activity preferentially toward PI(4,5)P₂. This 5-phosphate phosphatase is highly expressed in brain and particularly abundant in presynaptic terminals. It is involved in the regulation of synaptic endocytosis. This localization is consistent with its pathogenic role in Down syndrome and the early onset of Alzheimer’s disease (Berman et al., 2008; McPherson et al., 1996; Voronov et al., 2008).

SHIP1 and SHIP2 are homologous 5’-phosphatases which can utilize PI(3,4,5)P₃ or PI(1,3,4,5)P₄ as their substrate. SHIP1 expression is restricted to cells of the hematopoietic lineage. Ship1−/− mice develop myeloproliferative syndrome (Helgason et al., 1998), a fact that has implications for immunity and leukemias (Kerr, 2011; Kerr et al., 2011). In contrast, studies of SHIP2 have linked its polymorphism to metabolic syndromes such as insulin resistance and hypertension (Kaisaki et al., 2004; Marcano et al., 2007; Marion et al., 2002). Taken together, diverse functions are expected from these 5’-phosphate enzymes. Detailed data regarding their substrate specificity, anatomical localization, and their interacting protein partners would be very helpful in deciphering these variable functions.

Conclusion

The deficiency of FIG4 in humans and mice likely results in a clinical and pathological phenotype that is consistent with a lysosomal storage disorder. This disorder profoundly affects spinal motor neurons and glia. These findings have thus uncovered a signaling pathway involving FIG4/PI3,5P₂ that appears to be required for the regulation of lysosomal functions. Because lysosomal dysfunction has been implicated in many neurodegenerative disorders and myelin-related diseases, Fig4/PI3,5P₂ signaling may have much broader implications than previously thought.

Despite the remarkable progress that has been made towards understanding the biology of Fig4, there are still many unanswered questions concerning the pathogenesis of Fig4 deficiency. Based on current literature, we propose a hypothetical mechanism in Figure 1B. This figure is by no means a complete story, but it should provide a simplified theoretical platform for future exploration. Early in the endolysosomal pathway, the PAS complex is assembled and localized to the endosomes prior to their maturation to multi-vesicular bodies (Ho et al., 2012; Sbrissa et al., 2008; Shisheva, 2012). PI(3,5)P₂ begins to be synthesized on endosomal membranes and is additionally required for the activation of lysosomal TRPML1/MCOLN1 channels (Dong et al., 2010; Shen et al., 2012). Thus, a deficiency of FIG4/PI(3,5)P₂ would impair TRPML1/MCOLN1 channel function, leading to the accumulation of calcium in the lysosomes. Alternatively, TRPML1/MCOLN1 might fail to be delivered to the lysosomal membrane secondarily to a mislocalization or expression of other lysosomal membrane proteins. This would decrease calcium release from lysosomes as well, but would occur independent of PI(3,5)P₂. These alterations are speculated to inhibit the fission of cytoplasmic membrane targeted vesicles out of lysosomes (Dove et al., 2009a; Zhang et al., 2008), which would cause lysosomal storage. In cells with a lower lysosomal pH, abnormal lysosomal storage with high electron density forms. In contrast, electron lucent vacuolated lysosomes would form in cells with a higher lysosomal pH (Kogot-Levin et al., 2009). Mechanisms responsible for variable lysosomal pH in different cells are unknown at this point. In addition, the abnormal lysosomal storage would impair the trafficking or degradation of cytoplasmic voltage-gated calcium channels (Kiselyov et al., 2010; Tsuruta et al., 2009), leading an increase of cytoplasmic calcium channels and intracellular calcium concentration. The increase of calcium in myelinating Schwann cells is known to cause segmental demyelination (Smith et al., 2001; Smith and Hall, 1988).

Highlights.

• Loss of function of FIG4 causes severe neuronal degeneration in humans and mice.

• The predominant pathology of FIG4 deficiency is abnormal lysosomal storage.

• We discuss how the abnormal lysosomal storage contributes to pathological changes.

Abbreviation List

PI(3,5)P₂

phosphatidylinositol 3,5-bisphosphate

CMT4J

Charcot-Marie-Tooth Disease type 4J

kD

kilodalton

PI3P

phosphatidylinositol 3-phosphate

plt

pale tremor

CNS

central nervous system

PNS

peripheral nervous system

DRG

dorsal root ganglion

TRPML1

transient receptor potential mucolipin-1

LEL

late endolysosome

EM

electron microscopy

OL

myelinating oligodendrocyte

OPC

oligodendrocyte precursor cell

ALS

amyotrophic lateral sclerosis