Abstract
Loss of function of FIG4 leads to Charcot-Marie-Tooth disease Type 4J, Yunis-Varon syndrome, or an epilepsy syndrome. FIG4 is a phosphatase with its catalytic specificity toward 5′-phosphate of phosphatidylinositol-3,5-diphosphate (PI3,5P2). However, the loss of FIG4 decreases PI3,5P2 levels likely due to FIG4's dominant effect in scaffolding a PI3,5P2 synthetic protein complex. At the cellular level, all these diseases share similar pathology with abnormal lysosomal storage and neuronal degeneration. Mice with no FIG4 expression (Fig4−/−) recapitulate the pathology in humans with FIG4 deficiency. Using a flow cytometry technique that rapidly quantifies lysosome sizes, we detected an impaired lysosomal fission, but normal fusion, in Fig4−/− cells. The fission defect was associated with a robust increase of intralysosomal Ca2+ in Fig4−/− cells, including FIG4-deficient neurons. This finding was consistent with a suppressed Ca2+ efflux of lysosomes because the endogenous ligand of lysosomal Ca2+ channel TRPML1 is PI3,5P2 that is deficient in Fig4−/− cells. We reactivated the TRPML1 channels by application of TRPML1 synthetic ligand, ML-SA1. This treatment reduced the intralysosomal Ca2+ level and rescued abnormal lysosomal storage in Fig4−/− culture cells and ex vivo DRGs. Furthermore, we found that the suppressed Ca2+ efflux in Fig4−/− culture cells and Fig4−/− mouse brains profoundly downregulated the expression/activity of dynamin-1, a GTPase known to scissor organelle membranes during fission. This downregulation made dynamin-1 unavailable for lysosomal fission. Together, our study revealed a novel mechanism explaining abnormal lysosomal storage in FIG4 deficiency. Synthetic ligands of the TRPML1 may become a potential therapy against diseases with FIG4 deficiency.
Introduction
Autosomal recessive mutations in the FIG4 gene cause three distinct neurological disorders: Charcot-Marie-Tooth disease Type 4J (CMT4J) with neuronal degeneration in the peripheral nervous system (Chow et al., 2007; Zhang et al., 2008; Nicholson et al., 2011), Yunis–Varon syndrome with mental retardation and cleidocranial dysplasia (Campeau et al., 2013), and seizures with cerebral polymicrogyria. FIG4 encodes a phosphatase with its catalytic specificity toward the 5′-phosphate of phosphatidylinositol-(3,5)-bisphosphate (PI3,5P2). FIG4 (SAC3 in mammalian cells) complexes with a scaffolding protein Vac14 (=ArPIKfyve) and a 5′-kinase of PI3P known as Fab1 (PIKfyve) (Jin et al., 2008; Ikonomov et al., 2009). This PAS complex mediates the conversion of early endosomal PI3P to late endosomal PI3,5P2 (Sbrissa et al., 2007; Huotari and Helenius, 2011). In this manner, FIG4 can decrease PI3,5P2 levels via its phosphatase action and also promote PI3,5P2 synthesis by acting as a secondary scaffold for the Fab1/Vac14 interaction. However, the latter function appears to be dominant as loss of FIG4 in Fig4−/− mice results in a reduction of PI3,5P2 (Chow et al., 2007).
Mice with homozygous truncation mutations in Fig4 (also called pale tremor = Fig4−/−) recapitulate the pathology seen in humans with FIG4 deficiency (Chow et al., 2007; Zhang et al., 2008; Katona et al., 2011). At the cellular level, all these diseases and Fig4−/− mice share a common feature: abnormal lysosomal storage (Walch et al., 2000; Katona et al., 2011; Martyn and Li, 2013). Although mechanisms leading to the lysosomal storage are unclear, excessive lysosomal storage in these diseases is consistent with the documented role of PI3,5P2 in endolysosomal membrane trafficking. This lysosomal storage is not related to abnormal autophagy (Katona et al., 2011; Ferguson et al., 2012; Martyn and Li, 2013). Therefore, unlike canonical lysosomal storage diseases with deficiency of enzymatic degradation, FIG4 deficiency affects a different aspect of lysosomal function, namely, membrane trafficking.
Lysosome size is dynamically regulated by lysosomal membrane fusion and fission (Durchfort et al., 2012). Equilibrium of the two processes would have to be tightly controlled to maintain proper lysosomal size. Molecular machinery executing fusion is relatively better understood. It initiates with tethering two vesicular membranes that require Rab GTPases. Rab recruits additional bridging proteins, such as the homotypic fusion and protein sorting complex or Vamp7, which promote a mixing of lipid bilayers for fusion (Luzio et al., 2007). In contrast, lysosomal fission is minimally explored. In the present study, we demonstrated an impaired lysosomal fission, but not fusion, in Fig4−/− cells. Moreover, we found that this fission defect is related to a robust increase of intralysosomal Ca2+ level in Fig4−/− cells. Release of the Ca2+ from Fig4−/− lysosomes rescued abnormal lysosomal storage.
Materials and Methods
Mouse genotyping, skin biopsies to culture fibroblasts, and production of conditionally immortalized Schwann cells.
The Fig4−/− mice were genotyped as described previously (Chow et al., 2007). Both males and females were used and matched between different experimental groups. Skin biopsies from Fig4+/+ and Fig4−/− mice were transferred to culture dishes. The tissues were submerged in DMEM (including 10% FBS and 1× penicillin/streptomycin) for fibroblast culture that will be detailed below.
The primary Schwann cell culture is time-consuming. We expected a large amount of Schwann cells needed in this and subsequent studies. We thus prepared conditionally immortalized Schwann cells. The property of primary cells was largely preserved in these immortalized cells (Cárdenas et al., 2002; Saavedra et al., 2008). Fig4+/− mice were crossed with SV40 transgenic mice (SV40tg; from The Jackson Laboratory) to produce Fig4+/−/SV40tg mice. At postnatal day 5 (P5), sciatic nerves were dissected and Schwann cells were cultured at 33°C. This low temperature activated SV40tg, which promoted cell proliferation to a large quantity (Cárdenas et al., 2002; Saavedra et al., 2008). Cells were allowed to differentiate by transferring them to a 37°C, which inactivated SV40tg and restored the primary property of these cells. Identity of Schwann cells was confirmed by staining the cells with S100 antibodies. Fig4−/−/SV40tg Schwann cells developed vacuoles, which were indistinguishable from those described in the original Fig4−/− fibroblasts (Chow et al., 2007; Zhang et al., 2008) (data not shown).
Cell culture and lysosomal labeling.
Rat Schwann cell-line (RT4) was purchased from (ATCC CRL-2768). Fibroblasts were isolated from Fig4+/+ and Fig4−/− mouse skin biopsies. Cells were seeded on 6-well plates and cultured in DMEM, supplemented with 10% FBS. Cells were incubated with FITC-dextran (5 μg/ml; D-3305, Invitrogen) overnight followed by 2 h chasing to label lysosomes. Acetate Ringer's medium (to fragment lysosomes) contained 80 mm NaCl, 70 mm sodium acetate, 5 mm KCl, 2 mm CaCl2, 1 mm MgCl2, 2 mm NaH2PO4, 10 mm HEPES, 10 mm glucose, and 0.5 mg/ml BSA.
Isolation of intracellular organelles.
Fibroblasts in 6-well plates grown until 80% confluence, followed by incubation with FITC-dextran to label lysosomes. Cells were washed twice with 4°C DMEM-no Phenol Red (Invitrogen) and transferred into 4°C FACS loading buffer (275 mm sucrose, 20 mm HEPES, pH 7.2, 1 mm EGTA). Cytoplasmic membranes were ruptured using a Dunce homogenizer (Kimble/Kontes; #885303-0002) to stroke up and down 20 times. Homogenates were centrifuged at 400 × g for 4 min to collect intracellular organelles from supernatants (Perou et al., 1997). Unruptured cells or nuclei were removed during centrifugation.
Flow cytometry.
Organelles were placed in a 5 ml polystyrene tube (BD Falcon; #352058) for cytometry on BD-LSRII at Vanderbilt Flow Cytometry Core. The analysis was done using software BD FACSDiva version 6.1.3 and Flowjo version 10. A sample of unlabeled organelles was used to set a background for isolation of the FITC-dextran-labeled lysosomes. Because a fraction of lysosomes were large in size, the upper limit of forward scatter area (FSC-A) values (representing organelle size) was open to include all large organelles. Nuclei and unruptured cells were removed by centrifugation. Multiple organelles could cluster into a single large particle. These particles were excluded by setting up an additional discrimination window based on BD FACService TechNotes (BD 9:4; October, 2004). Typically, there were 10,000 events collected for each sample. To reduce variations between experiments, each measurement was performed with a standard sample of liposomes with a diameter of 400 nm. All FSC-A values were normalized by the FSC-A value of the 400 nm liposomes.
Confocal imaging.
Schwann cells, fibroblasts, or neurons were cultured on 35 mm glass-bottom dishes (In Vitro Scientific; #D35-14-1.5-N). Cells were preloaded with the FITC-Dextran and/or transfected with red fluorescence protein (RFP)-LAMP1 vector (CellLight Lysosomes-RFP, BacMam 2.0; Molecular Probes) and grown overnight. Fluorescence was imaged under a Zeiss LSM510 confocal microscopy at Vanderbilt Imaging Core.
For measuring Ca2+ levels, we used two different dyes: Calcium Orange (4 μm, Molecular Probes; #C3015) or Oregon Green (10 μm; Molecular Probes; catalog #06809). The cells were then stained with Calcium Orange for 45 min or Oregon Green for 2.5 h before the dye was removed. Cells were incubated for another 2 h to allow the Oregon Green to enter the intracellular organelles before imaging. Calcium Orange was imaged right after the dye was washed out with nonphenol red DMEM.
Ex vivo preparation of DRG.
Spinal cord segments with attached DRGs were dissected from P5 Fig4−/− mice. Each segment was cut in a half at the mid-line with one-half for vehicle and the other for ML-SA1 (40 μm) treatment for 36 h incubation in neurobasal media. The tissues were then fixed in 4% PFA for 24 h before being embedded into paraffin blocks. Paraffin sections were stained with cresyl violet (0.2% in acetate buffer) for 1–2 min, rinsed in 70% ethanol, and followed by nucleus staining with DAPI. Under fluorescent microscopy, DRG neuronal cell body was visualized through the autofluorescence of cresyl violet. Intracellular vacuoles with a diameter ≥2 μm were readily detectable due to the absence of autofluorescence. We counted percentages of neurons with vacuoles (diameters ≥2 μm) as described previously (Katona et al., 2011).
Morphological analysis of lysosomal vacuoles.
In a subset of fibroblast culture experiments, lysosome size was quantified manually under a 63× objective lens. We imaged five predefined fields on each slide. Images were imported into ImageJ. The areas of each field were measured. Within each field, we counted fibroblasts with large lysosomal vacuoles that had their diameters more than one-third of the nuclear diameter.
GTPase activity assay.
This method was modified from a previous publication (Zhang et al., 2013). Cells were washed twice with ice-cold PBS and lysed in 0.3 ml buffer (Mg2+ lysis/wash buffer + 1× protease inhibitor mixture) and centrifugation at 13,000 × g for 5 min. Supernatant were incubated with 0.25 volume of GTP-agarose (Sigma; G9768) overnight at 4°C on a rotator. Beads were washed and centrifuged in the lysis buffers. GTP-binding proteins were eluted with an equal volume of elution buffer (5–10 mm GTPγS in the lysis buffer) for 2 h at 4°C. The supernatant was subjected to Western blot.
Western blot.
Cells were lysed in RIPA buffer (#R0278, Sigma) with proteinase/phosphatase inhibitor mixture (#5872S, Cell Signaling Technology). The protein concentration was determined by BCA protein assay (#23225, Thermo Scientific). The protein sample was loaded into SDS-polyacrylamide gel. The remaining steps were similar to what we described (Katona et al., 2011).
Patch-clamp recording on Fig4+/+ and Fig4−/− Schwann cells.
Conditionally immortalized Schwann cells were dissociated by brief exposure to trypsin, plated on glass coverslips, and allowed to recover for ∼1 h at 37°C in 5% CO2. Whole-cell currents were recorded at room temperature (20°C–23°C) using Axopatch 200B amplifiers (Molecular Devices) in the whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). Pulse generation was performed via Clampex 8.0 (Molecular Devices) and whole-cell currents were filtered at 1 kHz and acquired at 5 kHz. The access resistance and apparent membrane capacitance were estimated using an established method (Lindau and Neher, 1988). Whole-cell currents were not leak-subtracted. Whole-cell currents were measured from −80 to 60 mV (in 20 mV steps) from a holding potential of 0 mV. We used solutions similar to those previously used (Mignen and Shuttleworth, 2001). The control external solution contained (in mm) the following: NaCl 140, MgCl2 1.2, CaCl2 20, glucose 10, HEPES 10, pH 7.4. The divalent-fee external solution contained the following (in mm): NaCl 140, glucose 10, HEPES 10, sucrose 30, pH 7.4. The internal solution contained the following (in mm): CsOH 140, aspartic acid 140, MgCl2 1.22, EGTA 5, HEPES 10, pH 7.2. Pipette solution was diluted 5%–10% to prevent activation of swelling-activated currents. Patch pipettes were pulled from thick-wall borosilicate glass (World Precision Instruments) with a multistage P-97 Flaming-Brown micropipette puller (Sutter Instruments) and heat-polished with a Microforge MF 830. After heat polishing, the resistance of the patch pipettes was 3–5 MΩ in the standard extracellular solution. The access resistance varied from 5 to 9 MΩ. As a reference electrode, a 2% agar-bridge with composition similar to the control bath solution was used. Junction potentials were zeroed with the filled pipette in the bath solution. Data were analyzed and plotted using a combination of Clampfit (Molecular Devices) and SigmaPlot 2000 (Systat Software). Whole-cell currents are normalized for membrane capacitance, and results are expressed as mean ± SEM.
Statistics.
We compared continuous variables between two groups using Student's t test. The difference among more than two groups was compared using one-way ANOVA if data were under normal distribution or Wilcoxon Mann–Whitney Test if data were not under normal distribution. p values of <0.05 indicated a statistical significance. Exceptions are listed below.
For Figure 2A, to compare slopes of changes over time for the two groups, we used a linear regression model with normalized FSC-A values as the outcome variable and group, time, and group × time as independent variables. Parameters from this model were estimated and then compared with test if the slopes were equal for the two groups. For Figure 2B, the same model was used, except by replacing outcome variable with fold of FSC-A value changes. The null hypothesis was H0: slope for Fig4+/+ = slope for Fig4−/− group. The slopes for the two groups were similar and showed no evidence to reject null hypothesis (p > 0.05). For Figure 2C, D, we compared the change values from baseline (or vehicle) between the two groups using Wilcoxon Rank-sum test. Exact p values were calculated because of the small sample sizes.
Results
Establishment of an automated quantification system for lysosome sizes
To quantitatively examine the fission of lysosomes, we established an automated technique that can rapidly measure the lysosome sizes by modifying a reported method (Durchfort et al., 2012) (Fig. 1). The method is detailed in Materials and Methods, including lysosomal labeling with FITC-conjugated dextran, intracellular organelle isolation, and flow cytometry analysis.
The RT4 Schwann cell line was treated with vacuolin-1 (5 μm/L), a reagent known to enlarge lysosomes by an unknown mechanism (Cerny et al., 2004). There was a significant increase of lysosomal sizes in vacuolin-1-treated cells (10.4 ± 1.2 FSC-A values; n = 12 cultures) compared with that in the vehicle-treated cells (5.6 ± 0.6; n = 11; p < 0.01; Fig. 1F,G,K). The increases of lysosomal sizes by vacuolin-1 were dose-dependent (Fig. 1H). Moreover, this experiment revealed significantly increased lysosome sizes in Fig4−/− mouse fibroblasts (10.1 ± 1.2; n = 6 cultures), compared with those of Fig4+/+ fibroblasts (6.8 ± 0.8; n = 8; p < 0.01; Fig. 1I–K). Therefore, this method is reliable, rapid, and can quantify thousands of lysosomes within a minute.
Lysosomal fission, but not fusion, is decreased in Fig4−/− cells
Next, we tested whether there is an altered balance between lysosomal fusion and fission in Fig4−/− cells. Mouse fibroblasts were incubated with Acetate Ringer's medium to fragment lysosomes (Perou and Kaplan, 1993; Durchfort et al., 2012). After the removal of Ringer's medium, lysosome sizes were tracked by flow cytometry to monitor their fusion activity (Fig. 2A). There was no significant difference in the rate of lysosome size increase between Fig4+/+ and Fig4−/− fibroblasts (Fig. 2A,B).
We then tested lysosomal fission. Fibroblasts were treated with vacuolin-1 to increase lysosome size (Fig. 2C). After removal of vacuolin-1, lysosome sizes were measured by flow cytometry to track their fission activity. By the 15th hour, the size of lysosomes reached an 83% recovery in Fig4+/+ cells, versus a 37% recovery in Fig4−/− cells.
One might be concerned about large lysosomes being fractured during homogenization of the cells, leading to an underestimation of lysosome sizes in Fig4−/− cells. We examined the live cells by DIC imaging and counted the percentage of cells with vacuole sizes more than one-third of the nuclear diameter. A strong recovery was evident in Fig4+/+ cells. However, lysosomes in Fig4−/− cells were further enlarged even after withdrawal of vacuolin-1 and showed no recovery (Fig. 2D).
Together, these results demonstrate that FIG4 deficiency selectively impairs lysosomal fission but spares lysosomal fusion. This mechanism explains abnormal lysosomal storage in Fig4−/− cells.
High Ca2+ levels in Fig4−/− lysosomes
To understand how lysosomal fission is impaired in Fig4−/− cells, we explored intralysosomal Ca2+ levels since this change has been involved in lysosomal storage (Lloyd-Evans and Platt, 2011). Moreover, PI3,5P2 is an endogenous ligand of the Ca2+ channel TRPML1 located in the lysosomal membrane (Dong et al., 2010). Deficiency of PI3,5P2 in Fig4−/− cells is expected to deactivate the TRPML1 channels and suppresses the Ca2+ release from Fig4−/− lysosomes. We stained the Fig4−/− fibroblasts with a Ca2+ dye (Calcium Orange; Fig. 3A,B, red). Ca2+ fluorescence intensity was significantly higher in Fig4−/− cells than that in Fig4+/+ fibroblasts (Fig. 3G).
Ca2+ signals were strong and often masked the dextran signals in Fig4−/− cells. There were stronger dextran signals in Fig4−/− cells than that in Fig4+/+ cells (Fig. 3B). This was consistent with excessive lysosomal storage in Fig4−/− cells. Robust Ca2+ signals were clearly visible in individual vacuoles of lysosome (Fig. 3B). Therefore, higher Ca2+ levels were the cause for the stronger signals in Fig4−/− lysosomes, rather than through an increase in the number of lysosomes. The increase of intralysosomal Ca2+ levels was also observed in Fig4−/− Schwann cell culture (Fig. 3C,D), Fig4−/− neuron culture (Fig. 3E,F,H), as well as in vivo myelinated Schwann cells of Fig4−/− mouse sciatic nerves (Fig. 3I–L).
Alteration of lysosomal pH may affect the results of the calcium dye. We thus tested the Ca2+ levels in Schwann cell cultures using pH-insensitive dye, Oregon Green 488 BAPTA-2-AM (10 μm; Molecular Probes, catalog #06809) (Takahashi et al., 1999). Again, the results demonstrated a significant increase of Ca2+ level in Fig4−/− cells (50.6 ± 19.9 gray values/μm2 in 51 Fig4−/− cells vs 24.1 ± 10.0 gray values/μm2 in 42 Fig4+/+ cells; p < 0.001).
Moreover, we have examined the Ca2+ sources in Schwann cells by using a culture medium either free of Ca2+ or added with 1.8 mm Ca2+. Schwann cells were incubated with media for 2 h. The cells were then stained with a calcium dye, Oregon Green, and imaged as described in Materials and Methods. Ca2+ fluorescent intensity was not significantly different between Fig4−/− cells with Ca2+-free medium and Fig4−/− cells with Ca2+-added medium (208.1 ± 62.6 gray values/μm2 for Ca2+-added medium vs 191.4 ± 51.5 gray values/μm2 for Ca2+-free medium; 70 cells from 20 fields under 63× lens or 100 cells from 19 fields under 63× lens respectively; p > 0.05). In addition, using the patch-clamp technique, Ca2+ influx across cytoplasmic membrane store-operated Ca2+ channels (conducting extracellular Ca2+ to intracellular endoplasmic reticulum) was evaluated in immortalized Schwann cells and showed no difference between Fig4+/+ and Fig4−/− cells (Fig. 4). Finally, we performed Western blot with pan-voltage-gated Ca2+ channel antibodies to compare the Ca2+ channel levels at cellular membrane between Fig4+/+ and Fig4−/− mouse sciatic nerves. Again, there was no difference between the two types of cells (data not shown).
To further substantiate the intralysosomal localization of Ca2+, we transfected Schwann cells using the BacMam vector 2.0 (Invitrogen; catalog #C10597) to express a red fluorescence protein (RFP)-fused LAMP1, a lysosomal transmembrane protein. This specifically labeled the outer margin of lysosomes (Fig. 3M). More importantly, this allowed us to clearly visualize the lysosomal lumen in the enlarged lysosomes of Fig4−/− cells. When these cells were double-labeled with the calcium dye, Oregon Green, Ca2+ were unequivocally localized within the lumen of lysosomes (Fig. 3M). Together, an abnormal Ca2+ increase was mainly in the lysosomes and was an abnormality common to all types of Fig4−/− cells.
High intralysosomal Ca2+ level and abnormal lysosomal storage are rescued by pharmacologically activating TRPML1 channels
The high intralysosomal Ca2+ level in Fig4−/− cells is highly relevant to abnormal lysosomal storage because autosomal recessive mutations in human TRPML1 gene cause a lysosomal storage disease, called mucolipidosis Type IV (Cox and Cachón-González, 2012). Thus, it is reasonable to speculate that the high level of intralysosomal Ca2+ by FIG4/PI3,5P2 deficiency can be reversed by application of a previously established synthetic TRPML1 ligand, ML-SA1 (Dong et al., 2010; Samie et al., 2013). Indeed, application of ML-SA1 reduced the level of intralysosomal Ca2+ (Fig. 5A–F).
Because ablation of Trpml1 in mice induced abnormal lysosomal storage (Venugopal et al., 2007), reactivation of TRPML1 by ML-SA1 is expected to improve lysosomal storage in Fig4−/− cells. In addition, the excessive Ca2+ may be released by activating another lysosomal Ca2+ channel, TPC2, through application of the TPC2 ligand, NAADP (Pitt et al., 2010). If the high Ca2+ level is pathogenic, release of the intralysosomal Ca2+ should alleviate the abnormal lysosomal storage in Fig4−/− cells. Lysosome sizes in mouse Fig4−/− fibroblasts were quantified by flow cytometry and compared between treated (20 μm ML-SA1 or 1 μm NAADP-AM; 24 h) and vehicle cells. Both ML-SA1 and NAADP-AM reduced lysosome sizes (Fig. 5G). NAADP cannot penetrate the cell membrane; therefore, NAADP was modified to NAADP-AM (acetoxylmethyl) by Vanderbilt Chemical Synthesis Core as described previously (Park et al., 2013). Acetoxylmethylation is a widely used chemical modification to render a compound membrane-permeable. Percentages of fibroblasts with large vacuoles were counted as described in Figure 2D to show the dose dependence response to ML-SA1 treatment (Fig. 5H).
Next, we tested ML-SA1's therapeutic potential in an ex vivo preparation (Fig. 5I,J). Spinal cords with attached DRGs from P5 mice (n = 4 ML-SA1-treated Fig4−/− mice and n = 4 vehicle-treated Fig4−/− mice) were treated with either vehicle (DMSO) or ML-SA1 (40 μm) for 36 h. We counted percentages of neurons with vacuoles (Fig. 5K). There was a significant decrease of vacuolated neurons (Fig. 5I,J, arrowheads) in ML-SA1-treated DRGs (12 ± 5%), compared with that in vehicle-treated DRGs (45 ± 3%; n = 4 mice; p < 0.01; 346 cells counted in vehicle group and 438 cells counted in ML-SA1 group).
These data together suggest that the high Ca2+ level in Fig4−/− cells is predominantly present in lysosomes. This likely resulted from deactivation of TRPML1 channels when PI3,5P2 is deficient. Activation of these TRPML1 Ca2+ channels pharmacologically rescues the abnormal lysosomal storage in both cell culture and ex vivo nervous tissues, supporting a pathogenic role of high Ca2+ levels in Fig4−/− lysosomes.
TRPML1 is expressed in the peripheral nervous system and increased in Fig4−/− fibroblasts
Lysosomal fission defect could also be contributed by a downregulation of TRPML1 (the receptor of PI3,5P2) in Fig4−/− cells. However, Western blot revealed an increase, not decrease, of TRPML1 protein level in Fig4−/− fibroblasts (0.73 ± 0.11) compared with that in Fig4+/+ fibroblasts (0.42 ± 0.04; n = 3 assays; p < 0.05; Fig. 6A,B). This increase is presumably due to an overall increase of lysosomal membrane proteins in Fig4−/− mice (Katona et al., 2011).
To delineate expression of TRPML1 in the nervous system, we performed immunostaining of mouse tissues with antibodies against TRPML1. TRPML1 was detected in the neuronal cell body of spinal cords (Fig. 6C). TRPML1 was also expressed in myelin but absent in axons of peripheral nerves (Fig. 6D, asterisk). The pattern of TRPML1 localization appeared to be comparable between Fig4−/− and Fig4+/+ nerve tissues.
Suppressed Ca2+ efflux from Fig4−/− lysosomes impairs lysosomal fission by downregulating dynamin-1 expression/activity
To understand how the high intralysosomal Ca2+ level affects lysosomal fission, we investigated dynamin, a calcium-sensitive GTPase that is known to play key roles in endoplasmic reticulum fission by scissoring the budding vesicles (Sweitzer and Hinshaw, 1998; Stowell et al., 1999). A recent study has demonstrated that dynamin is required for lysosomal fission (Schulze et al., 2013). Because dynamin is activated by calcium, we hypothesize that the lysosomal fission defect in Fig4−/− cells is due to impaired dynamin function.
Downregulation of dynamin-1 expression/activity in Fig4−/− cells
Immunostaining was used to detect dynamin-1 in Fig4+/+ and Fig4−/− fibroblasts. There appeared to be higher levels of dynamin-1 in Fig4+/+ cytoplasm than in Fig4−/− cytoplasm (Fig. 7A,B). This difference was confirmed by Western blot in the vehicle lanes of Figure 7C and the input lanes of Figure 7E (0.28 ± 0.07 in Fig4−/− cells vs 0.85 ± 0.08 in Fig4+/+ cells; n = 5 assays; p < 0.01).
To determine whether the difference of dynamin-1 expression is associated with a difference of dynamin-1 GTPase activity, we performed a GTPase activity assay by modifying a published method (Zhang et al., 2013). GTPase activity was present in Fig4+/+ fibroblasts but hardly detectable in Fig4−/− fibroblasts (Fig. 7E). However, the ratio between activated GTPase level and total dynamin-1 level was not significantly different between Fig4+/+ (5.6 ± 1.1%) and Fig4−/− cells (5.9 ± 1.0%; n = 3 assays; p = 0.69). Therefore, the reduction of dynamin-1 activity in Fig4−/− cells was mainly attributed to the decrease of total dynamin-1 expression.
Failure of expressing dynamin-1 in Fig4−/− cells during the high demand of lysosomal fission
Vacuolin-1 maximizes lysosome size through an unknown mechanism, which would demand highly active fission to restore the lysosomal sizes after the removal of vacuolin-1. We speculated that vacuolin-1 will raise the level of dynamin-1 in Fig4+/+ cells, but Fig4−/− cells may not respond to the vacuolin-1 stimuli. Fig4+/+ fibroblasts were treated with vacuolin-1 (5 or 10 μm for 1 h) to enlarge lysosomes. Cell lysates were collected for Western blot at 4 h after the removal of vacuolin-1 when lysosomal fission became highly active (Fig. 2D). There was a significant increase of dynamin-1 levels in Fig4+/+ cells. However, this increase of dynamin-1 was not seen in Fig4−/− cells after treatment with the vacuolin-1 (Fig. 7C,D).
To further determine the role of dynamin-1 in lysosomal fission, we have used dynasore, a specific dynamin-GTPase inhibitor (Fig. 8). Following a 1 h treatment of 10 μm vacuolin-1, lysosomes in wild-type fibroblasts reached their peak size around the 24th hour. After this point, lysosomal fission activity was expected to be high. Dynasore was added at the 25th hour to block lysosomal fission. All cells were imaged at the 69th hour and counted as described in Figure 2D. The percentage of cells with large vacuoles was 26 ± 4% in dynasore-treated fibroblasts but was only 10 ± 2% in vehicle-treated fibroblasts (n = 3 experiments; >50 cell counted in each experiment; p < 0.01; Fig. 8D). These data suggest that dynamin activity regulates fission of wild-type lysosomes.
The reduced dynamin-1 level is reversed by reactivation of lysosomal TRPML1 Ca2+ channel
We speculated that the reduction of dynamin-1 expression in Fig4−/− cells is related to the inactivation of Ca2+ efflux through lysosomal TRPML1 channels. We treated Fig4−/− fibroblasts with or without ML-SA1, followed by Western blot to quantify the dynamin-1 level. The results showed a significant increase of dynamin-1 in cells treated with ML-SA1 (1.7 ± 0.2), compared with that in vehicle-treated cells (1.1 ± 0.2; n = 6; p < 0.01; Fig. 9A,B). This finding suggests that suppressed Ca2+ efflux from lysosomes is one of the mechanisms responsible for the decreased dynamin-1 expression in Fig4−/− cells.
Dynamin-1 expression is also decreased in Fig4−/− mouse brain
To determine whether the mechanism of dynamin-1 reduction is pathogenically relevant in vivo, we have performed Western blot in Fig4+/+ and Fig4−/− mouse brains at P21 d of age. There was a significant difference of dynamin-1 level between Fig4−/− and Fig4+/+ mouse brains (4.2 ± 0.2; n = 3 Fig4+/+ mice vs 2.5 ± 0.5; n = 3 Fig4−/− mice; p < 0.01; Fig. 7G,H). The reduction of dynamin-1 could be secondary to the neuronal loss in Fig4−/− brains. To exclude this possibility, dynamin-1 levels were normalized by the neuronal marker PGP9.5. There was still a significant decrease of dynamin-1 in Fig4−/− brains, compared with that in Fig4+/+ brains. Moreover, the difference of dynamin-1 levels was replicated in P5 mouse brains, when neurodegeneration starts (3.7 ± 0.2 for 3 Fig4+/+ mouse brains vs 2.7 ± 0.2 for 3 Fig4−/− mouse brains; p < 0.01).
Together, data both in vitro and in vivo suggest that inactivation of TRPML1 Ca2+ channels downregulates dynamin-1 expression, thereby leading to the impaired fission and abnormal storage of lysosomes.
Discussion
Although PI3,5P2 has been proposed to be involved in lysosomal membrane trafficking for a long time, it is still unknown how PI3,5P2 deficiency affects lysosomal trafficking in diseases with FIG4 deficiency. Our study demonstrates a fission defect of lysosomes in Fig4−/− cells. This finding explains abnormal lysosomal storage in FIG4-deficient human cells and mouse model (Chow et al., 2007; Zhang et al., 2008; Katona et al., 2011; Martyn and Li, 2013). This novel mechanism is in line with our time-lapse imaging study showing a decrease of budding vesicles in FIG4-deficient human fibroblasts (Zhang et al., 2008).
We further show that defective lysosomal fission is likely caused by a robust increase of intralysosomal Ca2+ level in Fig4−/− cells. This mechanism is consistent with the known biology of TRPML1. This hexa-span membrane protein forms Ca2+ channels in the lysosomal membrane with both N- and C-terminals in the cytoplasmic side. Several positively charged residues in the N-terminal of TRPML1 allow cytoplasmic PI3,5P2 to bind with TRPML1 with high specificity. This binding promotes Ca2+ efflux through the TRPML1 channels as demonstrated in an elegant patch-clamp study by Dong et al. (2010). Deficiency of PI3,5P2 in FIG4 deficiency would suppress Ca2+ efflux from lysosomes. Indeed, application of a synthetic ligand of TRPML1, ML-SA1, rescues lysosomal fission defect and abnormal lysosomal storage in Fig4−/− cells. This is also supported by the abnormal lysosomal storage in mice with null of Trpml1 (Venugopal et al., 2007). Finally, release of intralysosomal Ca2+ by activating a different lysosomal Ca2+ channel, TPC2, improved lysosomal storage as well. This finding further supports a causal role of intralysosomal Ca2+ in abnormal lysosomal storage. This approach is therapeutically significant because application of ML-SA1 reverses lysosomal pathology in ex vivo tissues from Fig4−/− mice (Fig. 5I–K).
Next, a logical question would be how the high intralysosomal Ca2+ level suppresses lysosomal fission in FIG4 deficiency. We provide evidence both in vitro and in vivo that the suppressed Ca2+ efflux through TRPML1 channels downregulates the expression of dynamin-1, a GTPase known to regulate fissions of intracellular organelles (Sweitzer and Hinshaw, 1998; Stowell et al., 1999). The finding is pathogenically relevant for the following reasons: (1) Dynamin was significantly decreased in Fig4−/− culture cells and Fig4−/− mouse brains. (2) The increased demand of lysosomal fission by Vacuolin-1 treatment was able to raise the dynamin level in Fig4+/+ cells but failed to do so in Fig4−/− cells. (3) Application of the dynamin-specific inhibitor Dynasore in Fig4+/+ cells blocked lysosomal fission. (4) The decrease of dynamin level was reversed by application of lysosomal Ca2+ channel TRPML1 activator, ML-SA1. Together, loss of the dynamin scissoring activity in FIG4-deficient cells offers satisfying explanations for the lysosomal fission defect in Fig4−/− cells. This finding is consistent with the fact that genetically deletion of dynamin-1 gene in mice results in drastically increased sizes of lysosomes (Schulze et al., 2013).
How Ca2+ activates dynamin-1 remains to be determined. However, calcineurin is a Ca2+-responsive protein that directly binds to dynamin (Cabeza et al., 2010). Ca2+ from lysosomes may be required to activate the local dynamin GTPase via calcineurin (Cousin and Robinson, 2000). It also remains to be determined how the efflux of intralysosomal Ca2+ regulates expression of dynamin-1.
The lysosome is a recently emerged Ca2+ signaling sensor of the cell (Lloyd-Evans and Platt, 2011). This is a remarkable conceptual advance because an increase in intralysosomal Ca2+ level has become a common mechanism shared by multiple lysosomal storage diseases (Lloyd-Evans and Platt, 2011; Shen et al., 2012). By targeting this process, an effective therapeutic strategy has been developed against Niemann–Pick disease Type C1 (Lloyd-Evans et al., 2008). Moreover, identification of TRPML1 ligands has provided specific chemical tools to manipulate this system. We now extend this approach to a group of diseases caused by FIG4 deficiency.
TRPML1 synthetic ligands have no binding receptor in mucolipidosis-IV because recessive mutations in TRPML1 result in no/minimal TRPML1 translated. Even if mutant TRPML1s are translated, they cannot bind with their ligand PI3,5P2 due to the receptor TRPML1 being mutated. Thus, a disease, such as mucolipidosis-IV, offers no therapeutic target for the TRPML1 synthetic ligands (Dong et al., 2010; Wakabayashi et al., 2011). In contrast, diseases with FIG4 deficiency serve as excellent models to test the biological functions of TRPML1/PI3,5P2 interaction in animals and humans. A reversal of lysosomal phenotype by ML-SA1 supports that PI3,5P2 interaction with TRPML1 is a predominant mechanism for abnormal lysosomal storage in FIG4-deficient cells.
Furthermore, the implication of this approach may go beyond FIG4 deficiency. For instance, deficiency of two PI3,5P2 phosphatases, MTMR2 or MTMR13, causes CMT4B1 and CMT4B2, respectively (Bolino et al., 2002). This phosphatase abnormality leads to an increase of PI3,5P2 levels (Vaccari et al., 2011), which would overstimulate TRPML1 channels in CMT4B1 or 4B2. Antagonists of TRPML1 may prove to be a therapeutic approach against these diseases.
Lysosomal storage could be caused by additional mechanisms, such as alteration of lysosomal pH or suppression of protein degradation. In addition, a recent publication shows an autophagic defect in Fig4−/− Schwann cells (Vaccari et al., 2015). However, in our hands, defect of autophagy was negligible and was controversial based on other studies (Zhou et al., 2010; Katona et al., 2011; Ferguson et al., 2012; Martyn and Li, 2013). More importantly, all these alternative mechanisms are not mutually exclusive with our proposed mechanism of TRPML1 Ca2+ channel suppression.
In conclusion (see Fig. 10), our results revealed a defect of lysosomal fission, but not fusion in FIG4-deficient cells. This defect appears to be caused by a suppression of lysosomal TRPML1 Ca2+ channel activity due to the deficiency of the TRPML1 ligand, PI3,5P2. The failure of Ca2+ release from Fig4−/− lysosomes downregulates the dynamin-1 expression. Because dynamin-1 is required for lysosomal membrane scissoring, deficiency of dynamin-1 in FIG4-deficient cells explains the lysosomal fission defect and abnormal storage of lysosomes. Together with phenotypic reversal by ML-SA1 in ex vivo neural tissues, our data offer a potential therapeutic strategy by using TRPML1 synthetic ligands in the diseases with FIG4 deficiency.
Footnotes
J.Z. is a PhD candidate who has been supported by Natural Science Foundation of China (81129019). This work was supported by National Institute of Neurological Disorders and Stroke Grants R01NS066927 and R21NS081364 to J.L. and Department of Veterans Affairs Research and Development funds to J.L. We thank the Vanderbilt Flow Cytometry Shared Resource and Vanderbilt Cell Imaging Shared Resource for technical support; and Dr. Lily Wang (Department of Biostatistics, Vanderbilt University) for assistance in our statistical analysis.
The authors declare no competing financial interests.
- Correspondence should be addressed to Dr. Jun Li, Department of Neurology, Vanderbilt University School of Medicine, 1161 21st Avenue South, Nashville, TN 37232. jun.li.2{at}vanderbilt.edu