RESEARCH ARTICLE

Impaired myelination and reduced brain ferric iron in the mousemodel of mucolipidosis IVYulia Grishchuk1,‡, Karina A. Pen ̃a

2, Jessica Coblentz2, Victoria E. King1, Daniel M. Humphrey1, Shirley L.Wang1,Kirill I. Kiselyov2,* and Susan A. Slaugenhaupt1,*

ABSTRACTMucolipidosis type IV (MLIV) is a lysosomal storage disease causedby mutations in the MCOLN1 gene, which encodes the lysosomaltransient receptor potential ion channel mucolipin-1 (TRPML1). MLIVcauses impaired motor and cognitive development, progressive lossof vision and gastric achlorhydria. How loss of TRPML1 leads tosevere psychomotor retardation is currently unknown, and there isno therapy for MLIV. White matter abnormalities and a hypoplasticcorpus callosum are the major hallmarks of MLIV brain pathology.Here, we report that loss of TRPML1 in mice results in developmentalaberrations of brain myelination as a result of deficient maturationand loss of oligodendrocytes. Defective myelination is evidentin Mcoln1−/− mice at postnatal day 10, an active stage ofpostnatal myelination in the mouse brain. Expression of matureoligodendrocyte markers is reduced in Mcoln1−/− mice at postnatalday 10 and remains lower throughout the course of the disease. Weobserved reduced Perls’ staining in Mcoln1−/− brain, indicating lowerlevels of ferric iron. Total iron content in unperfused brain is notsignificantly different between Mcoln1−/− and wild-type littermatemice, suggesting that the observed maturation delay or lossof oligodendrocytes might be caused by impaired iron handling,rather than by global iron deficiency. Overall, these data emphasizea developmental rather than a degenerative disease course inMLIV, and suggest that there should be a stronger focus onoligodendrocyte maturation and survival to better understand MLIVpathogenesis and aid treatment development.

KEY WORDS: Lysosome, Myelination, Oligodendrocytes, Transientreceptor potential channel mucolipin-1

INTRODUCTIONMucolipidosis type IV (MLIV) was originally described in 1974based on clinical manifestations that resembled othermucolipidoses, but with a distinct biochemical profile (Bermanet al., 1974). In 2000, the gene mutated in MLIV was identified asMCOLN1, which encodes the protein mucolipin-1 (Bargal et al.,2000; Bassi et al., 2000; Sun et al., 2000). Based on homology withtransient receptor potential (TRP) ion-channel proteins, mucolipin-1 is classified as a founding member of the TRPML family and is

also known as TRPML1. Owing to its localization in the lysosomesand the observed lysosomal storage disease phenotype of MLIV,TRPML1 is presumed to regulate membrane traffic, lysosomalbiogenesis or lysosomal ion content (Colletti and Kiselyov, 2011;Wang et al., 2014). The diagnostic ultrastructural hallmark of MLIVis the accumulation of storage bodies filled with mixed lamellarand granular electron-dense content, called ‘compound’ bodies(Bargal and Bach, 1989; Berman et al., 1974). The biochemicalcomposition of this storage material includes gangliosides,phospholipids and acidic mucopolysaccharides (Bach et al., 1975,1977; Bach and Desnick, 1988; Bagdon et al., 1989; Bargal andBach, 1988, 1989; Slaugenhaupt et al., 1989).

MLIV is primarily found in the Ashkenazi Jewish population(approximately 75% of cases), with a carrier frequency of ∼1:100(Altarescu et al., 2002; Bach, 2001). There are approximately 100diagnosed individuals in the USA. Worldwide numbers are largelyunknown, but affected individuals with Arab (Mirabelli-Badenieret al., 2014), African-American (Noffke et al., 2001), FrenchCanadian (Berman et al., 1974) and Turkish descent (Tuysuz et al.,2009) have been reported. Over two dozen mutations causingMLIVhave been described, the vast majority of which lead to mRNAinstability and complete loss of protein (Wakabayashi et al., 2011).A small fraction of MLIV-causing mutations, some of which mightbe associated with a milder phenotype, lead to aberrant localizationand/or ion-channel dysfunction (Chen et al., 2014; Kiselyov et al.,2005). Establishing a genotype-phenotype correlation in MLIV iscomplicated because of the low incidence of the disease.

MLIV is a developmental disorder that causes motor andcognitive deficiencies, which become noticeable during the firstyear of life. On average, maximal neurological development inMLIV-affected individuals corresponds to the level of 15-18 months of age and remains stable throughout the third decadeof life (Altarescu et al., 2002). In most affected individuals,neurological symptoms include spasticity, hypotonia, an inability towalk independently, ptosis, myopathic facies, drooling, difficultiesin chewing and swallowing, and severely impaired fine-motorfunction (Altarescu et al., 2002). Ophthalmic symptoms areprogressive and result in blindness in the second decade of life(Riedel et al., 1985; Smith et al., 2002; Wakabayashi et al., 2011).

Human brain pathology data in MLIV are limited to only twoautopsy cases, in which gliosis, abnormal white matter andnumerous storage inclusions in all types of brain cells have beenreported (Folkerth et al., 1995; Tellez-Nagel et al., 1976). Brainmagnetic resonance imaging studies revealed dysgenic corpuscallosum, impaired myelination in the white matter and decreasedT2 signal intensities in the thalamus as a result of increased ferritindeposition (Frei et al., 1998; Schiffmann et al., 2014).

At present there is no therapy for MLIV, and one of the centralfactors hindering the search for treatments is a limited understandingof MLIV brain pathology. Mouse models of human disease provideReceived 10 April 2015; Accepted 8 September 2015

1Center for Human Genetic Research and Department of Neurology,Massachusetts General Hospital and Harvard Medical School, 185 CambridgeStreet, Boston, MA 02114, USA. 2Department of Biological Sciences, University ofPittsburgh, 519 Langley Hall, 4249 Fifth Avenue, Pittsburgh, PA 15260, USA.*These authors are co-senior authors

‡Author for correspondence (ygrishchuk@mgh.harvard.edu)

This is an Open Access article distributed under the terms of the Creative Commons AttributionLicense (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use,distribution and reproduction in any medium provided that the original work is properly attributed.

1591

© 2015. Published by The Company of Biologists Ltd | Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

an opportunity to gain a deeper insight into disease pathogenesisand progression. TheMcoln1 knockout mouse is an excellent modelof the human disease, because all hallmarks of MLIV are presentin these mice, with the exception of corneal clouding (Venugopalet al., 2007). MLIV mice display brain and gastric phenotypesidentical to those reported in humans affected with MLIV (Chandraet al., 2011; Grishchuk et al., 2014). Mcoln1−/− animals display noimmediate phenotype at birth. At 2 months of age, they start todisplay motor and cognitive deficits that can be detected in the openfield test (Grishchuk et al., 2014). Motor deficits in MLIV miceprogress to total hindlimb paralysis and death by 8-9 months of age(Venugopal et al., 2007). Post-mortem brain analysis showslysosomal inclusions in multiple cell types, including neurons,astrocytes, oligodendrocytes, microglia and endothelial cells.Defective myelination and gliosis, the hallmarks of human MLIV,are present in the young adult mice at the onset of cognitive andmotor deficits (Grishchuk et al., 2014). However, we observed no

neuronal loss in the Mcoln1−/− forebrain even at the late stage ofdisease (Grishchuk et al., 2014). These findings promptedinvestigation of glial cell involvement in MLIV. In the presentstudy, we use the MLIV mouse model to analyze the dynamics ofmyelination and oligodendrocyte development in the MLIV mouseand to investigate the role of iron dyshomeostasis in MLIVpathogenesis. We show abnormal oligodendrocyte developmentand reduced ferric iron levels in the white matter of the Mcoln1−/−

mice. These findings highlight the role of TRPML1 in proper brainmyelination and suggest that there should be a stronger focus onoligodendrocyte maturation and survival in the search for potentialtherapeutic approaches in MLIV.

RESULTSImmunostaining of brain coronal sections with antibodies againstmature myelinating oligodendrocyte marker proteolipid protein(PLP; Baumann and Pham-Dinh, 2001; Han et al., 2013) showedreduced myelination in Mcoln1−/− mice (Fig. 1). We observed asignificant reduction in the PLP immunoreactivity measured in thewhole hemispheric coronal section or in the cortical region in adultmice (Fig. 1A,B). We also observed a significant reduction in thesurface area of the myelinated corpus callosum. Both findings werepresent in the early-symptomatic (2 months of age) and the late stageMcoln1−/− mice (7 months of age), but were not progressive in thecourse of disease. Similar data were obtained on the brain sectionsfrom 2-month-old Mcoln1−/− and wild-type littermate mice,immunostained with the antibodies against myelin basic protein(MBP; Fig. S1; Huang et al., 2013). This striking pattern of reducedmyelination and thinned corpus callosum is similar to the previouslyreported FluoroMyelin Green staining performed in Mcoln1−/−

mice at 2 and 7 months of age (Grishchuk et al., 2014). Reduced andnon-progressive myelination inMcoln1−/−mice was also confirmedby western blotting, which showed a fourfold decrease in the levelsof PLP and MBP in the cerebral cortex at the age of 2 and 7 months(Fig. 1C,D). Interestingly, myelination deficits (shown as reducedlevels of MBP) were also detected outside the central nervoussystem, in sciatic nerves of Mcoln1−/− mice (Fig. S2).

In order to determine whether reduced myelination in the whiteand gray matter of the Mcoln1−/− brain is secondary to a lowerdensity of axons, we performed staining with the pan-axonal markerSMI312 (Vasung et al., 2011). Strikingly, we observed no changesin the surface area of the SMI312-stained corpus callosum and in theSMI312 immunoreactivity in the whole hemi-brain sections,cortical region or corpus callosum between genotypes at 2 and7 months of age (Fig. 2A,B). These data show that white matterabnormalities and reduced gray matter myelination are not the resultof a reduced density of axonal fibers in Mcoln1−/− mice.

In order to determine whether the loss of TRPML1 alters thedevelopmental stage of myelin deposition, we analyzed myelinationmarkers at postnatal day 10, an active phase of brain myelination inmice. PLP immunostaining of brain sections showed a disorganizedpattern of myelinating fibers in the corpus callosum and in thecortex of Mcoln1−/− pups (Fig. 3A), although the overall intensitywas not significantly different between genotypes at thisdevelopmental stage. Quantitative analysis of fiber morphologyin the corpus callosum showed a significant reduction in thenumber of branches, number of junctions and number of triple andquadruple points in Mcoln1−/− pups (Table 1 and Fig. S3),reflecting decelerated development of the myelinated networkowing to loss of TRPML1. Moreover, western blot analysis ofcortical homogenates showed reduced levels of PLP and MBP inMcoln1−/−mice (2.5- and threefold, respectively; Fig. 3B). Overall,

TRANSLATIONAL IMPACT

Clinical issue

Mucolipidosis type IV (MLIV) is a lysosomal storage disease caused bymutations in MCOLN1, which encodes the lysosomal ion channelTRPML1 (mucolipin-1). The disease is characterized by delayed motorand cognitive development, which becomes noticeable during the firstyear of life. Neurological development in individuals with MLIVcorresponds on average to that of an unaffected individual at 15months of age and remains stable in the second and third decades oflife. White-matter abnormalities and a hypoplastic corpus callosum arethe major hallmarks of brain pathology in MLIV. At present, however,pathogenesis of MLIV is poorly understood and thus there is no specifictreatment for this disease.

Results

The authors show that loss of TRPML1 in the mouse model of MLIV(Mcoln1 knockout mouse) results in reduced myelination in both thewhite-matter tracts and the gray matter of the brain, similar to what isobserved in MLIV human brains. Furthermore, their data indicate thatoligodendrocyte differentiation, maturation and region-dependent loss ofoligodendrocytes contribute to dysmyelination in Mcoln1−/− mice. Iron isa key factor required for oligodendrocyte development and myelination,and iron deficiency in children is often associated with hypomyelinationand neurological disabilities. The authors observed reduced Perls’staining in the Mcoln1−/− brain, indicating lowered levels of ferric iron,although the total iron content in the brain was not changed. Thesefindings indicate that dysmyelination in MLIV could be caused byimpaired iron handling in the brain. Mechanistically, the role of TRPML1in iron trafficking in the brain could be explained either by its function as alysosomal iron channel that releases ferrous iron from the lysosome tothe cytoplasm, or by its involvement in the endolysosomal membranetrafficking pathways that are important for iron delivery and distributionwithin the brain.

Implications and future directions

Iron is known to be vital for brain development, yet mechanisms of itsregulation are poorly understood. This study provides evidence thatTRPML1 might be an important player, with implications forunderstanding the molecular basis of MLIV. Future work will befocused on determining the mechanism underlying dysmyelination inMLIV, with the aim of devising new therapeutic strategies for thisincurable, devastating disease. Beyond this rare disease, a betterunderstanding of the role of TRPML1 in brain iron regulation provides anew target for therapies aimed at overcoming the consequences ofchildhood brain iron deficiency, the most common micronutrientdeficiency worldwide.

1592

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

these data indicate early developmental changes in brainmyelination as a result of loss of TRPML1.The observed reduction of myelination resulting from the loss of

TRPML1 could reflect the following factors: (1) the loss ofdifferentiated oligodendrocytes or their abnormal differentiation;(2) deficient myelin formation by Mcoln1−/− oligodendrocytes; or(3) a combination of both. In order to determine whether

reduced myelination in the Mcoln1−/− brain is caused by reducednumbers of oligodendrocytes, we performed immunostaining withadenomatous polyposis coli (APC-CC1), a marker for postmitoticpremyelinating oligodendrocytes, in 10-day-old and 2-month-oldmice. Analysis of CC1-positive (CC1+) cell density revealed thatthe number of oligodendrocytes is significantly reduced in theMcoln1−/− corpus callosum at 10 days of age, but the CC1+ cell

Fig. 1. Reduced myelination in Mcoln1−/− mouse brain. (A,B) Proteolipid protein (PLP) immunostaining on coronal brain hemi-sections from 2- (A) and 7-month-old (B)Mcoln1−/−mice and control littermates shows reducedmean pixel intensity in both the hemi-section and cortical region, and reduced surface area ofcorpus callosum inMcoln1−/− sections. Right image panels represent magnified images of the regions of interest shown in thewhite boxed areas in the left panels.Scale bars: left panels=500 µm; right panels=100 µm. n=5 per genotype (2 months old); n=3 per genotype (7 months old). Bar graphs on the right show statisticalanalysis of mean pixel intensities and surface areas obtained using these sample groups. (C,D) Western blot images and quantification show a significantreduction in the level of PLP and myelin basic protein (MBP) in cerebral cortex homogenates fromMcoln1−/− mice at the age of 2 (C) and 7 months (D). Samplesfrom three individual mice per genotype group are loaded in duplicates. All graphs represent average values ±s.e.m. Student’s t-test: *P<0.05, n.s. P>0.05.

1593

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

density improved by 2 months of age (Fig. 4A). Interestingly, in thecortical region the CC1+ cell density was the same at 10 days of age,but a loss of cortical oligodendrocytes was evident by 2months. Theobserved difference between the corpus callosum and the cortexsuggests that there are region-specific oligodendrocyte responses in

the Mcoln1−/− brain, which could be caused by differences in themicroenvironment in these brain structures or by cell-intrinsicdifferences between white- and gray-matter oligodendrocytes.Although the mechanism is largely unknown, differences inoligodendrocyte differentiation and survival, as well as

Fig. 2. Preserved axonal projections in Mcoln1−/− mouse brain. Immunostaining with the pan-axonal marker SMI312 of coronal brain hemi-sections from2- (A) and 7-month-old (B)Mcoln1−/−mice and control littermates shows no significant difference in SMI312mean pixel intensity or surface area of corpora callosabetween control and Mcoln1−/− sections, except for a slight increase in SMI312 mean pixel intensity in the corpus callosum of 7-month-old Mcoln1−/− mice.All graphs represent average values ±s.e.m. Student’s t-test: *P<0.05, n.s. P>0.05.

Fig. 3. Myelin disruption in 10-day-old Mcoln1−/− mice. (A) Proteolipid protein (PLP) immunostaining on coronal brain hemi-sections shows morphologicallyaltered, ‘patchy’myelination inMcoln1−/− mice compared with wild-type littermates, whereas no significant changes in the overall mean pixel intensity or surfacearea were observed. Right panels are magnified images of the regions of interest shown in the white boxed areas of the left panels. Scale bars: leftpanels=500 µm; right panels=100 µm. Bar graphs on the right show statistical analysis of mean pixel intensities and surface areas obtained using these samplegroups. (B) Western blot images and quantification showing significant reduction in the levels of PLP and myelin basic protein (MBP) in cerebral cortex ofMcoln1−/− pups. Samples from three individual mice per genotype group are loaded in duplicates. All graphs represent average values ±s.e.m., n=3 per genotype.Student’s t-test: *P<0.05, n.s. P>0.05.

1594

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

differences in pathology in multiple sclerosis lesions between whiteand gray matter have been recently reported (Hill et al., 2014, 2013;Stadelmann et al., 2008).Next, we performed qRT-PCR gene-expression analysis using

whole-brain RNA extracts from 10-day-old, 2- and 7-month-oldmice. We observed reduced levels of mRNA encoding the mature

oligodendrocyte markers myelin-associated glycoprotein (MAG)and myelin-associated oligodendrocyte basic protein (MOBP; Hanet al., 2013) in Mcoln1−/− mice at 10 days (Fig. 5A). Significantdecline in mRNA levels of Mag and Mobp persisted in Mcoln1−/−

mice at 2 and 7 months of age (Fig. 5B), but was not progressive inthe course of disease as shown by a significant effect of genotype(Mag, Fgenotype=32.96, P<0.0001; Mobp, Fgenotype=23.67,P=0.0004) and no significant effect of age (Mag, Fage=0.12,P=0.73; Mobp, Fage=0.19, P=0.67) by two-way ANOVA. Thesedata provide further confirmation that deficient myelination inMcoln1−/− brain occurs during development and persists throughoutthe course of disease.

In order to determine whether impaired differentiation ofoligodendrocyte precursors contributes to a decreased densityof postmitotic oligodendrocytes and hypomyelination in theMcoln1−/− brain, we analyzed expression of the oligodendrocyteprecursor marker PDGFRα (Han et al., 2013; Perez et al., 2013)in Mcoln1−/− and wild-type mice at 10 days, 2 and 7 monthsof age. We observed a significant reduction in Pdgfra mRNAlevels in 10-day-old Mcoln1−/− mice, and no significantdifferences between genotypes in adult mice when myelination islargely complete (Fig. 5A,B). Lower levels of Pdgfra transcriptsat 10 days of age might indicate a reduced number of

Fig. 4. Decreased density ofoligodendrocytes in Mcoln1−/− brain.(A) Adenomatous polyposis coli(APC-CC1) immunostaining in the corpuscallosum (outlined by white dotted lines) of10-day- and 2-month-old Mcoln1−/− miceand their wild-type littermates. Scale bars:50 µm. Two-way ANOVA of CC1-positive(CC1+) cell (postmitotic oligodendrocyte)density shows a significant effect of age(F=47.11; P<0.0001) and genotype(F=12.34; P=0.0034). CC1+ cell densitydecreased significantly in Mcoln1−/−

corpora callosa at 10 days of age(Bonferroni post hoc test, P<0.01).Postnatal day 10, n=4 per genotype;2 months old, n=5 per genotype. (B) APC-CC1 immunostaining in cerebral cortex of10-day- and 2 month-old Mcoln1−/− miceand their wild-type littermates. Scale bars:50 µm. Two-way ANOVA of CC1+ celldensity shows a significant interactionbetween age and genotype (F=14.74;P=0.0024). CC1+ cell density decreasedsignificantly in Mcoln1−/− cortex at2 months of age (Bonferroni post hoc test,P<0.01). Postnatal day 10, n=3 pergenotype; 2 months old, n=5 per genotype.

Table 1. Morphological analysis of corpus callosum in 10-day-oldMcoln1−/− mice

Counts (±s.e.m.)Wild type(n=5)

Mcoln1−/−

(n=4)

P-value(Student’st-test)

Number of branches 162.8 (±15.4) 108 (±11.1) 0.029Number of junctions 76.8 (±8.9) 39 (±7.4) 0.016Number of end-point voxels 86.4 (±9.6) 109 (±15.9) 0.242Number of junction voxels 176.2 (±21.7) 87.75 (±18.3) 0.019Number of slab voxels 835.6 (±43.9) 541.75 (±49.9) 0.003Average branch length 138 (±28.5) 206.1 (±40.95) 0.201Number of triple points 59.2 (±6.2) 32.25 (±6.02) 0.018Number of quadruple points 14 (±2.1) 5 (±0.82) 0.009Maximal branch length 199.48 (±40) 295.7 (±58.65) 0.199

Bold indicates statistically significant changes between genotypes, P(Student’s t-test)<0.05.Average branch length and maximal branch length are expressed in pixels.

1595

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

oligodendrocyte precursors in Mcoln1−/− brain at this critical stageof myelination, perhaps limiting their differentiation, maturationand myelin deposition. Restoration of the Pdgfra mRNA levels inadult Mcoln1 knockout mice is not accompanied by restoration ofthe mature oligodendrocyte or myelin markers, highlighting theimportance of the time-dependent regulation of this process andperhaps multiple roles of TRPML1 in the oligodendrocyte celllineage.Given that iron is an important factor for oligodendrocyte

differentiation and maturation (Todorich et al., 2009) and thatTRPML1 is directly involved in intracellular iron trafficking(Coblentz et al., 2014; Dong et al., 2008), we set out to test theiron status in the Mcoln1−/− brain. We stained brain sections withPerls’ stain, which detects ferric iron bound to ferritin (Meguroet al., 2005). As the most iron-rich cells in the brain, matureoligodendrocytes are frequently visualized using this technique(Schonberg and McTigue, 2009). Mcoln1+/+ brains containedclearly identifiable Perls’ stain, but such staining was significantlyreduced in white-matter tracts inMcoln1−/− brains, indicating a lossof H-ferritin-bound iron (Fig. 6A-C). Measurement of total ironcontent (Fe2+ and Fe3+) in the unperfused brain by inductivelycoupled plasma mass spectrometry (ICP-MS) showed nostatistically significant differences in total iron levels between

Mcoln1−/− and wild-type mice, which rules out systemic irondeficiency in Mcoln1−/− mice (Fig. 6D). These data show that lossofMcoln1 results in iron dyshomeostasis in the brain, which is likelyto play a role in the observed impairment of brain myelination inMLIV.

Iron dyshomeostasis is known to lead to oxidative stress in thebrain (Abeliovich and Flint Beal, 2006; Chen et al., 2013; Hareet al., 2013; Schrag et al., 2011). Therefore, we evaluated the levelsof the marker Hmox1 mRNA in the brain of Mcoln1−/− and wild-type mice. Importantly,Hmox1mRNAwas significantly elevated inMcoln1−/− brain at all ages evaluated, indicating prolongedoxidative stress (Fig. 6E).

Overall, our data suggest that myelination deficits in Mcoln1−/−

mice are caused by defects in the formation of the myelin sheath,rather than by axonal loss. Our demonstration that markers ofoligodendrocyte precursors, premyelinating and mature myelinatingoligodendrocytes are all disrupted in Mcoln1−/− mouse brainsuggests that loss of TRPML1 affects all stages of oligodendrocytedevelopment. Furthermore, our results show impaired iron handlingin the brain as a result of loss of TRPML1. Given that iron is animportant trophic factor for oligodendrocyte differentiation andmaturation, improper iron handling as a result of loss of TRPML1might account for the observed dysmyelination in mucolipidosis IV.Our results, although descriptive in nature, provide the first evidenceof impaired iron handling and early myelination defects in MLIV,highlight the developmental nature of mucolipidosis IV andemphasize the rationale for a stronger focus on oligodendrogliaand early-stage myelination as potential targets for the developmentof therapeutics.

DISCUSSIONIn the present study, we show that hypomyelination in the MLIVmouse brain is associated with decreased expression of bothprecursor and mature oligodendrocyte markers, as well as a decreasein the number of postmitotic oligodendrocytes. Our observation ofdelayed myelin deposition during early postnatal brain developmenthighlights the developmental character of brain pathology in MLIV.Interestingly, our data also indicate that the fate of developingoligodendrocytes is affected differently in the white (corpuscallosum) versus gray (neocortex) matter of the brain as indicatedby changes in CC+ cell counts (Fig. 4) or PLP and MBP stainingintensities in these brain regions. This suggests that the tissuemicroenvironment and factors released by neighboring cells cancontribute to the ability of Mcoln1−/− oligodendrocytes to develop,survive and/or produce myelin. Despite these region-specificchanges within the brain, our data demonstrate that TRPML1 isan important regulator of myelination, both in the central nervoussystem and in peripheral nerves. Given the absence of overtneuronal loss in MLIV (Grishchuk et al., 2014), combined with thepreservation of axonal fibers (Fig. 2) in the Mcoln1−/− brain,oligodendrocyte development and survival should be high-prioritytargets for the development of therapeutics.

Themain function of oligodendrocytes is myelin synthesis and theformation of myelin sheaths around axons to facilitate actionpotential propagation. Pediatric disorders of myelin, such asperiventricular white matter injury and hereditary leukodystrophies,result in cerebral palsy-like and cognitive dysfunction similar to thosedocumented in MLIV (Pouwels et al., 2014). Proliferation ofoligodendrocyte precursors, differentiation and maturation relies onseveral trophic factors, including FGF, PDGF, IGF-1, thyroidhormone and iron (Morath and Mayer-Proschel, 2001; Rouault,2013; Todorich et al., 2009). Oligodendrocytes are themost iron-rich

Fig. 5. Decreased markers of oligodendrocyte precursors and matureoligodendrocytes in Mcoln1−/− mouse brain. qRT-PCR analysis of Pdgfra,Mag andMobp expression in 10-day-old (A), 2- and 7-month-old (B)Mcoln1−/−

and wild-type (WT) brains (n=4 per genotype for each time point). RNA of eachsample was extracted from whole brain. qRT-PCR reactions were performedblind to the investigator using numerically coded samples. Each reaction wasperformed in duplicate. The data were analyzed using the comparative ΔCtmethod and presented as a percentage of WT values. Student’s t-test wasused for statistical analysis of postnatal day 10 data; two-way ANOVA withBonferroni correction was used to assess the effect of genotype and age onmRNA levels for 2- and 7-month-old data. **P<0.01, *P<0.05, n.s. P>0.05.

1596

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

cells in the brain. As the most metabolically active cells in the brain,oligodendrocytes require massive amounts of ATP, and iron isinvolved in oxidative phosphorylation. Iron is a cofactor for a numberof enzymes directly or indirectly involved in myelin synthesis,including lipid saturase and desaturase (Nakamura and Nara, 2004;Rao et al., 1983; Todorich et al., 2009). Furthermore, iron deficiencycauses a delay in oligodendrocytematuration and hypomyelination inboth humans and rodents (Hare et al., 2013; Todorich et al., 2009).Injection of ferritin-bound iron into white matter leads to anincrease in the levels of oligodendrocyte progenitor and matureoligodendrocyte markers (Schonberg et al., 2012; Schonberg andMcTigue, 2009). An increase in intracellular ferritin levels is alsoevident during the oligodendrocyte maturation process (Li et al.,2013).TRPML1 is clearly required for oligodendrocyte development

and function, and this is perhaps what underlies the observed brainpathology in MLIV. The precise role of TRPML1 inoligodendrocyte development is currently unknown; however,several studies highlight possibilities and suggest a link to the roleof TRPML1 in iron handling. How might mucolipin-1 regulateiron handling? One possibility is through its function inendolysosomal trafficking. Mucolipin-1 localizes to lateendosomes and lysosomes (Kiselyov et al., 2005; Miedel et al.,2006; Vergarajauregui and Puertollano, 2006) and is known toregulate endolysosomal fusion and membrane trafficking (Donget al., 2010; LaPlante et al., 2004; Pryor et al., 2006; Samie et al.,

2013; Thompson et al., 2007; Vergarajauregui et al., 2008). Inmost tissues, iron is absorbed by endocytosis of Fe3+-transferrin orFe3+-ferritin (Valerio, 2007). In the endolysosomal compartment,Fe3+ is liberated from the protein complex, converted to Fe2+ andabsorbed into the cytoplasm via endolysosomal membranes. In thecytoplasm, Fe2+ is oxidized to Fe3+ and binds to cytoplasmicferritin. From the cytoplasm, iron enters mitochondria for energyproduction, binds to transferrin for trafficking or exporting,serves as a cofactor for enzymatic reactions, such as myelinproduction, or binds to cytoplasmic ferritin and other scavengersfor storage. Hence, based on the role of mucolipin-1 in endocyticmembrane trafficking, it might regulate Fe3+-ferritin delivery tothe lysosomes.

The second possibility is based on the permeability of TRPML1to divalent heavy metal ions, including Fe2+ (Dong et al., 2008), soit could facilitate the direct transport of Fe2+ from lysosome tocytoplasm. The common transferrin-dependent iron entry pathwayrelies on the divalent transporter DMT1 (Siah et al., 2006) for ironrelease from endolysosomes into the cytosol. Interestingly,oligodendrocytes depend strictly on extracellular ferritin as asource of iron (Hulet et al., 2000; Todorich et al., 2009, 2011,2008), and the role of DMT1 in iron handling specifically inoligodendrocytes has been disputed (Burdo et al., 2001; Song et al.,2007). Thus, owing to its Fe2+ permeability, mucolipin-1 mightsupplant DMT1 in oligodendrocytes totally, or during some stagesof development, thus making oligodendrocytes especially

Fig. 6. Decreased levels of ferric iron andoxidative stress in the Mcoln1−/− brain.(A) Coronal brain sections stained for Fe3+

content by modified Perls’ protocol. Note anincrease of iron deposits with age in the whitematter of control mice and reduced staining inthe Mcoln1−/− brain sections. Scale bar: 1 mm.(B) Higher-magnification images of Perls’-stained wild-type (WT) and Mcoln1−/− brainsections from white boxed areas shown in(A) demonstrate lower staining density inMcoln1−/− white matter. Scale bar: 50 µm.(C) Quantification of Perls’ staining showssignificantly reduced staining in Mcoln1−/−

brains at both 2 and 7 months of age (Student’st-test: P=0.025 and P=0.02, respectively); WT,n=5; Mcoln1−/−, n=4 at 2 months, and WT, n=4Mcoln1−/−, n=5 per genotype at 7 months.(D) Inductively coupled plasma massspectrometry (ICP-MS) analysis of brain cortexhomogenates of 7-month-old Mcoln1+/+ andMcoln1−/− mice done in a blind experiment (n=4per genotype). The differences are notsignificant (P=0.52 and P=0.30). (E) Oxidativestress in mucolipidosis type IV brain. qRT-PCRanalysis of Hmox1 expression in whole-brainhomogenates of Mcoln1+/+ and Mcoln1−/− mice(n=4 per genotype). Actb was used as ahousekeeping gene. The experiments wereperformed blind to investigator. The data arepresented as a percentage of WT mRNAlevels ±s.e.m. Student’s t-test was used forcomparison between genotypes for dataobtained in 10-day-old mice; *P<0.05. Two-wayANOVA of data sets at 2 and 7 months showed asignificant increase of Hmox1 expression withage (Fgenotype×age=7.2; P=0.014); Bonferronipost hoc test: ***P<0.0001, n.s. P>0.05.

1597

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

vulnerable to the loss of mucolipin-1. Our demonstration ofreduced levels of ferritin-bound ferric iron in Mcoln1−/− brain byPerls’ stain provides experimental support for this idea. Iron boundto ferritin, specifically to H-ferritin, is prevalent inoligodendrocytes and can be visualized using Perls’ stain.Reduced Perls’ staining in white-matter tracts of Mcoln1−/− miceindicates loss of biologically available, H-ferritin-bound iron.Given that other brain cell types, including neurons, rely ontransferrin endocytosis and DMT1 for iron delivery (Hare et al.,2013; Moos and Morgan, 2004; Paz Soldan and Pirko, 2012; Perezet al., 2013; Todorich et al., 2011; Valerio, 2007), it is possible thatloss of mucolipin-1 might not affect neuronal iron handling or totaliron content. Supporting this hypothesis, our ICP-MS data show nochanges betweenMcoln1−/− and control mice in the total amount ofiron in the unperfused brain tissue. Therefore, the maturation ormyelination deficits in MLIV cannot be explained by a generalbrain iron deficiency. Rather, they are directly related tooligodendrocyte malfunction.Given that TRPML1 is permeable to Fe2+ but not to Fe3+, it is

possible that loss of TRPML1 leads to lysosomal build-up ofFe2+, which is known to catalyze Fenton-like reactions thatresult in the formation of reactive oxygen species. Previously,oxidative stress has been shown both in in vitro cellularmodels (Coblentz et al., 2014) and in a Drosophila model(Venkatachalam et al., 2008), and in the present study we show,for the first time, evidence of oxidative stress in vivo in the MLIVmouse brain.Regardless of the precise role of the ion channel TRPML1 in

regulating iron in the brain, our studies suggest that it is a new targetfor pharmacological intervention for conditions linked to brain ironhandling and myelination. Our data also indicate that restoration ofmyelin is an attractive therapeutic strategy for MLIV, and furtherstudies will reveal whether mechanisms leading to impairedmyelination in MLIV rely solely on oligodendrocyte intrinsic(cell-autonomous) mechanisms.

MATERIALS AND METHODSAnimalsMcoln1 knockout mice were maintained and genotyped as previouslydescribed (Venugopal et al., 2007). The Mcoln1+/− breeders for this studywere obtained by backcrossing onto a C57Bl6J background for more than 10generations.Mcoln1+/+ littermates were used as controls. Experiments wereperformed according to the institutional and USNational Institutes of Healthguidelines and approved by the Massachusetts General HospitalInstitutional Animal Care and Use Committee.

Immunohistochemistry and image analysisTo obtain brain tissue for histological examination, 10-day-old, 2- and7-month-old Mcoln1−/− and control mice were transcardially perfusedunder isoflurane anesthesia with ice-cold PBS followed by 4%paraformaldehyde in PBS. Brains were postfixed in 4%paraformaldehyde in PBS for 24 h, washed with PBS, cryoprotected in30% sucrose in PBS overnight, frozen in isopentane and stored at −80°C.Brains were bisected along the midline, and one hemisphere wasexamined histologically. Coronal sections 40 μm thick were cut using aMicrom freezing microtome and collected into 96-well plates containingTBSAF (TBS, 30% ethylene glycol, 15% sucrose and 0.05% sodiumazide). These sections were stored at 4°C prior to any staining procedures.For experiments, the corresponding serial sections were used for eachbrain. For PLP, MBP and SMI312 staining, the sections were microwavedin citrate buffer (10 mM citric acid and 0.05% Tween 20, pH 6.0) at lowpower for 10 min for antigen retrieval, blocked in 0.5% Triton X-100 and5% normal goat serum (NGS) in PBS and incubated with primaryantibodies diluted in 1% NGS overnight. The following primary

antibodies were used: PLP rabbit polyclonal, 1:500, ab28486; MBP(SMI99), mouse monoclonal, 1:1000, ab24568; SMI312 (pan axonalneurofilament) mouse monoclonal 1:1000, ab24574; all from Abcam(Cambridge, MA, USA). Sections were incubated with secondaryantibodies in 1% NGS for 1.5 h at room temperature. The followingsecondary antibodies were used: goat anti-mouse AlexaFluor 488 anddonkey anti-rabbit AlexaFluor 555 (1:500; Invitrogen, Eugene, OR,USA). For APC-CC1 staining, no antigen retrieval was performed. Afterblocking, the sections were incubated with anti-APC-CC1 antibody(mouse monoclonal, 1:100, OP80; Calbiochem, Billerica, MA, USA),followed by goat anti-mouse AlexaFluor 633 (1:500; Invitrogen, Eugene,OR, USA). Sections were counterstained with NucBlue nuclear stain(Life Technologies, Eugene, OR, USA) and mounted on microscopicslides.

For PLP, MBP and SMI312 staining, images for analysis were acquiredon a Nikon 80i upright epifluorescence microscope with Hamamatsu OrcaCCD camera (Nikon, Tokyo, Japan) with 5× objective and NIS-Elements4.2 software (Nikon, Tokyo, Japan) using an automated stitching function.The exposure time was set to avoid saturation and was the same for allsections (wild-type and Mcoln1−/− sections within the sameimmunohistochemistry experiment). Image analysis was performed usingFIJI software (NIH, Bethesda, MD, USA). Areas of interest (whole hemi-section, cortical region or corpus callosum) were selected in each section.Area and mean pixel intensity values were compared between genotypesusing GraphPad Prizm 5 software (GraphPad, La Jolla, CA, USA) usingStudent’s t-test.

Quantitative morphological analysis of corpus callosum was done usingthe AnalyzeSkeleton plugin of ImageJ (Arganda-Carreras et al., 2010).Regions of interests (ROI; 200 pixels ×200 pixels) of corpus callosum wereselected on images of PLP-stained sections of 10-day-old brains. Selectionswere converted to 8-bit images, made binary and skeletonized using theSkeletonize 2D/3D option and analyzed using the AnalyzeSkeleton plugin.An investigator was blinded to genotypes. Total counts of number ofbranches, junctions, end-point voxels, junction voxels, slab voxels, tripleand quadruple points, average and maximal branch lengths per ROI permouse were averaged across animals in genotype groups (n=5 WT and n=4Mcoln1−/−) and compared in GraphPad Prizm 5 software using Student’st-test.

For APC-CC1 analysis, we acquired images of two non-overlappingfields of view in somatosensory or motor cortex and corpus callosumusing a 20× objective (HCX PL APO CS 20.0×0.70 DRY UV) andconfocal laser scanning microscope Leica TCS SP5 (Leica MicrosystemsInc., Wetzlar, Germany). The number of CC1+ cells was counted in FIJIin corresponding areas of interest (cortex or corpus callosum). CC1 celldensity was expressed as the number of cells per square millimeter of thetissue. Statistical analysis of data (two-way ANOVA with Bonferronicorrection to assess the effect of genotype and age) was performed usingGraphPad Prizm 5.

Perls’ stainingFor Perls’ staining, the sections were incubated in potassium ferrocyanidewith HCl [10% potassium ferrocyanide:10% HCl (2:1)] for 60 min withrocking. Sections were washed with water, incubated with inactivated 3,3′-diaminobenzidine (DAB) (no H2O2 added) and activated DAB, for 30 mineach time. After washing in water, sections were mounted on microscopeslides, dried and coverslipped. Bright-field images were acquired using aLeica DMI6000 microscope using DFC425 camera and HCX PL S-APO5.0×0.15 objective (Leica Microsystems Inc.). Inverted mean pixelintensities were measured in FIJI. For background normalization, invertedmean pixel intensity measured outside the brain tissue was subtracted frominverted mean pixel intensity values for every image. Student’s t-test wasused for comparisons between genotypes.

Western blot analysisFor immunoblotting, fresh-frozen pieces of sensory-motor cortex werehomogenized in the lysis buffer (20 mmol/l HEPES, pH 7.4, 10 mmol/lNaCl, 3 mmol/l MgCl2, 2.5 mmol/l EGTA, 0.1 mmol/l dithiothreitol,50 mmol/l NaF, 1 mmol/l Na3VO4 and 1% Triton X-100; reagents from

1598

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

Sigma, LaJolla, CA, USA) and a protease inhibitor cocktail (catalog number11873580001; Roche, Mannheim, Germany). Homogenates werecentrifuged at 2100 g for 10 min at 4°C, supernatants collected andprotein concentration was determined using Pierce BSA Protein Assay kit(ThermoFisher Scientific, Waltham, MA, USA). Sciatic nerves werehomogenized in radioimmunoprecipitation assay buffer (RIPA; BostonBioProducts, Ashland, MA, USA) supplemented with 0.1 mmol/l EDTAand a protease inhibitor cocktail (catalog number 11873580001; Roche,Mannheim, Germany) according to Vargas et al. (2010). Proteins (20 μg forcortex and sciatic nerve homogenates from 2- and 7-month-old mice; 40 μgfor 10-day-old cortex homogenates) were separated by SDS-PAGE on a 4-12% Bis-Tris gel 3-(N-morpholino)propansulfonic acid (MOPS) buffer(Invitrogen, Waltham, MA, USA) and transferred onto nitrocellulosemembrane (ThermoFisher Scientific). The following primary antibodieswere used: anti-proteolipid protein (ab28486, 1:1000, rabbit polyclonal;Abcam, Cambridge, MA, USA), anti-myelin basic protein (NE1019, mousemonoclonal, 1:1000; EMD Millipore, Billerica, MA, USA), anti-β-actin(A5441, 1:3000, mouse monoclonal; Sigma). After incubation with primaryantibody, the following secondary antibodies were applied: polyclonal goatanti-mouse or goat anti-rabbit IgG conjugated with IRDye 680 or IRDye800, both from LI-COR (Lincoln, NE, USA). Protein bands were visualizedusing the Odyssey Infrared Imaging System and analyzed by Odyssey v3.0software (LI-COR). In densitometric analysis, optical density values werenormalized by β-actin. All data were expressed as mean±s.e.m. Data wereanalyzed statistically using Student’s t-test.

RNA extraction and qRT-PCRFor qRT-PCR experiments, total RNAwas extracted from brain tissue from10-day-old, 2- and 7-month-old Mcoln1−/− and control mice using TRizol(Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s protocol.For complementary DNA synthesis, MuLV Reverse Transcriptase (AppliedBiosystems, Foster City, CA, USA) was used with 2 μg of total RNA andoligo(dT)18 (IDT, Cralville, IA, USA) as the primer. qRT-PCR was carriedout using 1:75 dilutions of complementary DNA, 2× SYBR Green/ROXMaster Mix (Fermentas, Glen Burnie, MD, USA), and 4 μM primer mix per10 μl reaction. For gene-expression analysis, the following primers wereused (IDT): Pdgfrα, forward 5′-TGGAAGCTTGGGGCTTACTTT-3′ andreverse 5′-TAGCTCCTGAGACCTTCTCCT-3′; Plp, forward 5′-GTTCC-AGAGGCCAACATCAA-3′ and reverse 5′-CCACAAACTTGTCGGGA-TGT-3′; Mbp, forward 5′-GTGCCACATGTACAAGGACT-3′ and reverse5′-TGGGTTTTCATCTTGGGTCC-3′; Mag, forward 5′-CGGGATTGTC-ACTGAGAGC-3′ and reverse 5′-AGGTCCAGTTCTGGGGATTC-3′;Mobp, forward 5′-CACCCTTCACCTTCCTCAAC-3′ and reverse 5′-TT-CTGGTAAAAGCAACCGCT-3′; Hmox1, forward 5′-CACAGATGGCG-TCACTTCCGTC-3′ and reverse 5′-GTGAGGACCCACTGGGAGGAG-3′; and Actb, forward 5′-GCTCCGGCATGTGCAAAG-3′ and reverse5′-CATCACACCCTGGTGCCTA-3′. To avoid amplification of genomicDNA and ensure amplification of complementary DNA, all primers weredesigned to span exons, and negative RT reactions (without reversetranscriptase) were performed as controls. Samples were amplified intriplicates on the 7300 Real Time System (Applied Biosystems) using thefollowing program: 2 min at 50°C, 10 min at 95°C, and 40 cycles at 95°C for15 s followed by 1 min at 60°C. In addition, a dissociation curve step wasrun to corroborate that amplification with specific primers resulted in oneproduct only (Fig. S4). The ΔΔCtmethod was used to calculate relative geneexpression, where Ct corresponds to the cycle threshold. ΔCt values werecalculated as the difference between Ct values from the target gene and thehousekeeping gene Actb (Table S1). Experiments were performed as adouble blind using numerically coded samples; the sample identity wasdisclosed at the final stage of data analysis. Data are represented as foldchange.

ICP-MSICP-MS analysis of mouse cortical tissues was performed at the Departmentof Geology and Planetary Science, University of Pittsburgh. Frozen brainsamples were coded and analyzed in a blinded manner. Mouse corticaltissues were weighed, dissolved in pure nitric acid, dried, reconstituted in2% nitric acid and analyzed using a PerkinElmer NexION 350 ICM-MS

Spectrometer (PerkinElmer Inc., Waltham, MA, USA). Iron concentrationwas normalized to the total wet tissueweight. In addition to iron, sodium andother metals were analyzed. As an additional analysis tool, ironconcentration was normalized to sodium content.

AcknowledgementsThe authors thank Drs James Connor and Amanda Snyder for help with the Perls’staining protocol. The authors thank Dr Daniel Bain (University of Pittsburgh) for helpwith ICP-MS.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsY.G. and K.I.K. designed the study. Y.G., K.A.P., J.C., V.E.K., D.M.H. and S.L.W.performed experiments and data analysis. Y.G., K.I.K. and S.A.S. wrote themanuscript.

FundingThis work was supported by grants from ML4 Foundation to S.A.S. and Y.G. andNational Institutes of Health grants HD058577 and ES01678 to K.I.K.

Supplementary informationSupplementary information available online athttp://dmm.biologists.org/lookup/suppl/doi:10.1242/dmm.021154/-/DC1

ReferencesAbeliovich, A. and Flint Beal, M. (2006). Parkinsonism genes: culprits and clues.

J. Neurochem. 99, 1062-1072.Altarescu, G., Sun, M., Moore, D. F., Smith, J. A., Wiggs, E. A., Solomon, B. I.,

Patronas, N. J., Frei, K. P., Gupta, S., Kaneski, C. R. et al. (2002). Theneurogenetics of mucolipidosis type IV. Neurology 59, 306-313.

Arganda-Carreras, I., Fernandez-Gonzalez, R., Munoz-Barrutia, A. and Ortiz-De-Solorzano, C. (2010). 3D reconstruction of histological sections: applicationto mammary gland tissue. Microsc. Res. Tech. 73, 1019-1029.

Bach, G. (2001). Mucolipidosis type IV. Mol. Genet. Metab. 73, 197-203.Bach, G. and Desnick, R. J. (1988). Lysosomal accumulation of phospholipids in

mucolipidosis IV cultured fibroblasts. Enzyme 40, 40-44.Bach, G., Cohen, M. M. and Kohn, G. (1975). Abnormal ganglioside accumulation

in cultured fibroblasts from patients with mucolipidosis IV.Biochem. Biophys. Res.Commun. 66, 1483-1490.

Bach, G., Ziegler, M., Kohn, G. and Cohen, M. M. (1977). Mucopolysaccharideaccumulation in cultured skin fibroblasts derived from patients with mucolipidosisIV. Am. J. Hum. Genet. 29, 610-618.

Bagdon, M. M., Lubs, M. L., Phillips, J. A., III, Murray, J. C., Slaugenhaupt, S. A.,Chakravarti, A., Adler, W. H. and Bach, G. (1989). Lysosomal accumulation ofphospholipids in mucolipidosis IV cultured fibroblasts. Am. J. Hum. Genet. 44,48-50.

Bargal, R. and Bach, G. (1988). Phospholipids accumulation in mucolipidosis IVcultured fibroblasts. J. Inherit. Metab. Dis. 11, 144-150.

Bargal, R. and Bach, G. (1989). Phosphatidylcholine storage in mucolipidosis IV.Clin. Chim. Acta 181, 167-174.

Bargal, R., Avidan, N., Ben-Asher, E., Olender, Z., Zeigler, M., Frumkin, A.,Raas-Rothschild, A., Glusman, G., Lancet, D. and Bach, G. (2000).Identification of the gene causing mucolipidosis type IV. Nat. Genet. 26,118-123.

Bassi, M. T., Manzoni, M., Monti, E., Pizzo, M. T., Ballabio, A. and Borsani, G.(2000). Cloning of the gene encoding a novel integral membrane protein,mucolipidin-and identification of the two major founder mutations causingmucolipidosis type IV. Am. J. Hum. Genet. 67, 1110-1120.

Baumann, N. and Pham-Dinh, D. (2001). Biology of oligodendrocyte and myelin inthe mammalian central nervous system. Physiol. Rev. 81, 871-927.

Berman, E. R., Livni, N., Shapira, E., Merin, S. and Levij, I. S. (1974). Congenitalcorneal clouding with abnormal systemic storage bodies: a new variant ofmucolipidosis. J. Pediatr. 84, 519-526.

Burdo, J. R., Menzies, S. L., Simpson, I. A., Garrick, L. M., Garrick, M. D., Dolan,K. G., Haile, D. J., Beard, J. L. and Connor, J. R. (2001). Distribution of divalentmetal transporter 1 and metal transport protein 1 in the normal and Belgrade rat.J. Neurosci. Res. 66, 1198-1207.

Chandra, M., Zhou, H., Li, Q., Muallem, S., Hofmann, S. L. and Soyombo, A. A.(2011). A role for the Ca2+ channel TRPML1 in gastric acid secretion, based onanalysis of knockout mice. Gastroenterology 140, 857-867.e1.

Chen, J., Marks, E., Lai, B., Zhang, Z., Duce, J. A., Lam, L. Q., Volitakis, I., Bush,A. I., Hersch, S. and Fox, J. H. (2013). Iron accumulates in Huntington’s diseaseneurons: protection by deferoxamine. PLoS ONE 8, e77023.

Chen, C.-C., Keller, M., Hess, M., Schiffmann, R., Urban, N., Wolfgardt, A.,Schaefer, M., Bracher, F., Biel, M., Wahl-Schott, C. et al. (2014). A small

1599

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

molecule restores function to TRPML1 mutant isoforms responsible formucolipidosis type IV. Nat. Commun. 5, 4681.

Coblentz, J., St Croix, C. and Kiselyov, K. (2014). Loss of TRPML1 promotesproduction of reactive oxygen species: is oxidative damage a factor inmucolipidosis type IV? Biochem. J. 457, 361-368.

Colletti, G. A. and Kiselyov, K. (2011). Trpml1. Adv. Exp. Med. Biol. 704, 209-219.Dong, X.-P., Cheng, X., Mills, E., Delling, M.,Wang, F., Kurz, T. and Xu, H. (2008).The type IV mucolipidosis-associated protein TRPML1 is an endolysosomal ironrelease channel. Nature 455, 992-996.

Dong, X.-P., Shen, D., Wang, X., Dawson, T., Li, X., Zhang, Q., Cheng, X., Zhang,Y., Weisman, L. S., Delling, M. et al. (2010). PI(3,5)P(2) controls membranetrafficking by direct activation of mucolipin Ca(2+) release channels in theendolysosome. Nat. Commun. 1, 38.

Folkerth, R. D., Alory, J., Lomakina, I., Skutelsky, E., Raghavan, S. S. andKolodny, E. H. (1995). Mucolipidosis IV: morphology and histochemistry of anautopsy case. J. Neuropathol. Exp. Neurol. 54, 154-164.

Frei, K. P., Patronas, N., Crutchfield, K. E., Altarescu, G. and Schiffmann, R.(1998). Mucolipidosis type IV: characteristic MRI findings. Neurology 51,565-569.

Grishchuk, Y., Sri, S., Rudinskiy, N., Ma,W., Stember, K. G., Cottle, M.W., Sapp,E., Difiglia, M., Muzikansky, A., Betensky, R. A. et al. (2014). Behavioraldeficits, early gliosis, dysmyelination and synaptic dysfunction in a mouse modelof mucolipidosis IV. Acta Neuropathol. Commun. 2, 133.

Han, H., Myllykoski, M., Ruskamo, S., Wang, C. and Kursula, P. (2013). Myelin-specific proteins: a structurally diverse group of membrane-interacting molecules.Biofactors 39, 233-241.

Hare, D., Ayton, S., Bush, A. and Lei, P. (2013). A delicate balance: ironmetabolism and diseases of the brain. Front. Aging Neurosci. 5, 34.

Hill, R. A., Patel, K. D., Medved, J., Reiss, A. M. and Nishiyama, A. (2013). NG2cells in white matter but not gray matter proliferate in response to PDGF.J. Neurosci. 33, 14558-14566.

Hill, R. A., Patel, K. D., Goncalves, C. M., Grutzendler, J. and Nishiyama, A.(2014). Modulation of oligodendrocyte generation during a critical temporalwindow after NG2 cell division. Nat. Neurosci. 17, 1518-1527.

Huang, H., Zhao, X.-F., Zheng, K. and Qiu, M. (2013). Regulation of the timing ofoligodendrocyte differentiation: mechanisms and perspectives.Neurosci. Bull. 29,155-164.

Hulet, S. W., Heyliger, S. O., Powers, S. and Connor, J. R. (2000).Oligodendrocyte progenitor cells internalize ferritin via clathrin-dependentreceptor mediated endocytosis. J. Neurosci. Res. 61, 52-60.

Kiselyov, K., Chen, J., Rbaibi, Y., Oberdick, D., Tjon-Kon-Sang, S.,Shcheynikov, N., Muallem, S. and Soyombo, A. (2005). TRP-ML1 is alysosomal monovalent cation channel that undergoes proteolytic cleavage.J. Biol. Chem. 280, 43218-43223.

LaPlante, J. M., Ye, C. P., Quinn, S. J., Goldin, E., Brown, E. M., Slaugenhaupt,S. A. and Vassilev, P. M. (2004). Functional links between mucolipin-1 and Ca2+-dependent membrane trafficking in mucolipidosis IV. Biochem. Biophys. Res.Commun. 322, 1384-1391.

Li, Y., Guan, Q., Chen, Y., Han, H., Liu, W. and Nie, Z. (2013). Transferrin receptorand ferritin-H are developmentally regulated in oligodendrocyte lineage cells.Neural Regen. Res. 8, 6-12.

Meguro, R., Asano, Y., Odagiri, S., Li, C., Iwatsuki, H. and Shoumura, K. (2005).The presence of ferric and ferrous iron in the nonheme iron store of residentmacrophages in different tissues and organs: histochemical demonstrations bythe perfusion-Perls and -Turnbull methods in the rat. Arch. Histol. Cytol. 68,171-183.

Miedel, M. T., Weixel, K. M., Bruns, J. R., Traub, L. M. and Weisz, O. A. (2006).Posttranslational cleavage and adaptor protein complex-dependent trafficking ofmucolipin-1. J. Biol. Chem. 281, 12751-12759.

Mirabelli-Badenier, M., Severino, M., Tappino, B., Tortora, D., Camia, F.,Zanaboni, C., Brera, F., Priolo, E., Rossi, A., Biancheri, R. et al. (2014). A novelhomozygous MCOLN1 double mutant allele leading to TRP channel domainablation underlies Mucolipidosis IV in an Italian Child. Metab. Brain Dis. 30,681-686.

Moos, T. and Morgan, E. H. (2004). The significance of the mutated divalent metaltransporter (DMT1) on iron transport into the Belgrade rat brain. J. Neurochem. 88,233-245.

Morath, D. J. andMayer-Proschel, M. (2001). Ironmodulates the differentiation of adistinct population of glial precursor cells into oligodendrocytes. Dev. Biol. 237,232-243.

Nakamura, M. T. and Nara, T. Y. (2004). Structure, function, and dietaryregulation of delta6, delta5, and delta9 desaturases. Annu. Rev. Nutr. 24,345-376.

Noffke, A. S., Feder, R. S., Greenwald, M. J., O’Grady, R. B. and Roth, S. I.(2001). Mucolipidosis IV in an African American patient with new findings onelectron microscopy. Cornea 20, 536-539.

Paz Soldan, M. M. and Pirko, I. (2012). Biogenesis and significance of centralnervous system myelin. Semin. Neurol. 32, 9-14.

Perez, M. J., Fernandez, N. and Pasquini, J. M. (2013). Oligodendrocytedifferentiation and signaling after transferrin internalization: a mechanism ofaction. Exp. Neurol. 248, 262-274.

Pouwels, P. J. W., Vanderver, A., Bernard, G., Wolf, N. I., Dreha-Kulczewksi,S. F., Deoni, S. C. L., Bertini, E., Kohlschutter, A., Richardson, W., Ffrench-Constant, C. et al. (2014). Hypomyelinating leukodystrophies: translationalresearch progress and prospects. Ann. Neurol. 76, 5-19.

Pryor, P. R., Reimann, F., Gribble, F. M. and Luzio, J. P. (2006). Mucolipin-1 is alysosomal membrane protein required for intracellular lactosylceramide traffic.Traffic 7, 1388-1398.

Rao, G. A., Crane, R. T. and Larkin, E. C. (1983). Reduction of hepaticstearoyl-CoA desaturase activity in rats fed iron-deficient diets. Lipids 18,573-575.

Riedel, K. G., Zwaan, J., Kenyon, K. R., Kolodny, E. H., Hanninen, L. and Albert,D. M. (1985). Ocular abnormalities in mucolipidosis IV. Am. J. Ophthalmol. 99,125-136.

Rouault, T. A. (2013). Iron metabolism in the CNS: implications forneurodegenerative diseases. Nat. Rev. Neurosci. 14, 551-564.

Samie, M., Wang, X., Zhang, X., Goschka, A., Li, X., Cheng, X., Gregg, E.,Azar, M., Zhuo, Y., Garrity, A. G. et al. (2013). A TRP channel in the lysosomeregulates large particle phagocytosis via focal exocytosis. Dev. Cell 26,511-524.

Schiffmann, R., Mayfield, J., Swift, C. and Nestrasil, I. (2014).Quantitative neuroimaging in mucolipidosis type IV. Mol. Genet. Metab. 111,147-151.

Schonberg, D. L. and McTigue, D. M. (2009). Iron is essential foroligodendrocyte genesis following intraspinal macrophage activation. Exp.Neurol. 218, 64-74.

Schonberg, D. L., Goldstein, E. Z., Sahinkaya, F. R., Wei, P., Popovich, P. G. andMcTigue, D. M. (2012). Ferritin stimulates oligodendrocyte genesis in the adultspinal cord and can be transferred from macrophages to NG2 cells in vivo.J. Neurosci. 32, 5374-5384.

Schrag, M., Mueller, C., Oyoyo, U., Smith, M. A. and Kirsch, W. M. (2011). Iron,zinc and copper in the Alzheimer’s disease brain: a quantitative meta-analysis.Some insight on the influence of citation bias on scientific opinion. Prog.Neurobiol. 94, 296-306.

Siah, C. W., Ombiga, J., Adams, L. A., Trinder, D. and Olynyk, J. K. (2006).Normal iron metabolism and the pathophysiology of iron overload disorders. Clin.Biochem. Rev. 27, 5-16.

Slaugenhaupt, S. A., Lewis, J. G., Chakravarti, A., Antonarakis, S. E. andBargal, R. (1989). Phosphatidylcholine storage in mucolipidosis IV. Genomics 4,579-591.

Smith, J. A., Chan, C.-C., Goldin, E. and Schiffmann, R. (2002). Noninvasivediagnosis and ophthalmic features of mucolipidosis type IV. Ophthalmology 109,588-594.

Song, N., Jiang, H.,Wang, J. and Xie, J.-X. (2007). Divalent metal transporter 1 up-regulation is involved in the 6-hydroxydopamine-induced ferrous iron influx.J. Neurosci. Res. 85, 3118-3126.

Stadelmann, C., Albert, M., Wegner, C. and Bruck, W. (2008). Cortical pathologyin multiple sclerosis. Curr. Opin. Neurol. 21, 229-234.

Sun, M., Goldin, E., Stahl, S., Falardeau, J. L., Kennedy, J. C., Acierno, J. S., Jr,Bove, C., Kaneski, C. R., Nagle, J., Bromley, M. C. et al. (2000). Mucolipidosistype IV is caused by mutations in a gene encoding a novel transient receptorpotential channel. Hum. Mol. Genet. 9, 2471-2478.

Tellez-Nagel, I, Rapin, I., Iwamoto, T., Johnson, A. B., Norton, W. T. andNitowsky, H. (1976). Mucolipidosis IV. Clinical, ultrastructural, histochemical,and chemical studies of a case, including a brain biopsy. Arch. Neurol. 33,828-835.

Thompson, E. G., Schaheen, L., Dang, H. and Fares, H. (2007). Lysosomaltrafficking functions of mucolipin-1 in murine macrophages. BMC Cell Biol. 8,54.

Todorich, B., Zhang, X., Slagle-Webb, B., Seaman, W. E. and Connor, J. R.(2008). Tim-2 is the receptor for H-ferritin on oligodendrocytes. J. Neurochem.107, 1495-1505.

Todorich, B., Pasquini, J. M., Garcia, C. I., Paez, P. M. and Connor, J. R. (2009).Oligodendrocytes and myelination: the role of iron. Glia 57, 467-478.

Todorich, B., Zhang, X. and Connor, J. R. (2011). H-ferritin is the major source ofiron for oligodendrocytes. Glia 59, 927-935.

Tuysuz, B., Goldin, E., Metin, B., Korkmaz, B. and Yalcinkaya, C. (2009).Mucolipidosis type IV in a Turkish boy associated with a novel MCOLN1mutation.Brain Dev. 31, 702-705.

Valerio, L. G. (2007). Mammalian iron metabolism. Toxicol. Mech. Methods 17,497-517.

Vargas, M. E., Watanabe, J., Singh, S. J., Robinson, W. H. and Barres, B. A.(2010). Endogenous antibodies promote rapid myelin clearance and effectiveaxon regeneration after nerve injury. Proc. Natl. Acad. Sci. USA 107,11993-11998.

Vasung, L., Jovanov-Milosevic, N., Pletikos, M., Mori, S., Judas, M. andKostovic, I. (2011). Prominent periventricular fiber system related to ganglionic

1600

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms

eminence and striatum in the human fetal cerebrum. Brain Struct. Funct. 215,237-253.

Venkatachalam, K., Long, A. A., Elsaesser, R., Nikolaeva, D., Broadie, K. andMontell, C. (2008). Motor deficit in a Drosophila model of mucolipidosis type IVdue to defective clearance of apoptotic cells. Cell 135, 838-851.

Venugopal, B., Browning, M. F., Curcio-Morelli, C., Varro, A., Michaud, N.,Nanthakumar, N., Walkley, S. U., Pickel, J. and Slaugenhaupt, S. A. (2007).Neurologic, gastric, and opthalmologic pathologies in a murine model ofmucolipidosis type IV. Am. J. Hum. Genet. 81, 1070-1083.

Vergarajauregui, S. and Puertollano, R. (2006). Two di-leucine motifs regulatetrafficking of mucolipin-1 to lysosomes. Traffic 7, 337-353.

Vergarajauregui, S., Connelly, P. S., Daniels, M. P. and Puertollano, R. (2008).Autophagic dysfunction in mucolipidosis type IV patients. Hum. Mol. Genet. 17,2723-2737.

Wakabayashi, K., Gustafson, A. M., Sidransky, E. and Goldin, E. (2011).Mucolipidosis type IV: an update. Mol. Genet. Metab. 104, 206-213.

Wang, W., Zhang, X., Gao, Q. and Xu, H. (2014). TRPML1: an ion channel in thelysosome. Handb. Exp. Pharmacol. 222, 631-645.

1601

RESEARCH ARTICLE Disease Models & Mechanisms (2015) 8, 1591-1601 doi:10.1242/dmm.021154

Disea

seModels&Mechan

isms