C: jag4 castillo.vp

J. Appl. Genet. 45(4), 2004, pp. 461-467 Tamoxifen and vitamin E treatments delay symptoms
in the mouse model of Niemann-Pick C
Eric C. BascuZan-Castillo1, Robert P. Erickson2, Christy M. Howison1, Robert J. Hunter2,Randall H. Heidenreich2, Chad Hicks3, Theodore P. Trouard4, Robert J. Gillies2,4 1 Department of Biochemistry and Molecular Biophysics, 2 Department of Pediatrics, 3 Optical Sciences Center, 4 BiomedicalEngineering Program, University of Arizona, Tucson, Arizona, USA Abstract. Niemann-Pick C disease (NPC) is an irreversible neurodegenerative disorder without current treat-
ment. It is the result of deficient intracellular cholesterol movement. We investigated the effects of tamoxifen and
vitamin E (D-alpha tocopherol) treatment on patterns of weight loss and motor function in the mouse model
of Niemann-Pick C disease (Npc1-/- mice). Tamoxifen has multiple metabolic effects, including reducing oxida-
tive damage, while vitamin E primarily has this property. Npc1-/- mice were identified and treatment was initi-
ated at an approximate age of 21 days. Tamoxifen suspended in peanut oil was administered via intraperitoneal
injection (weekly, at a dose calculated to deliver 0.023 ìg/g/day). Vitamin E (25 IU) was administered orally via
gavage once a week. Weight loss and Rota-Rod performance were analyzed by using Kaplan-Meyer survival
curves. Tamoxifen treatment by itself significantly delayed weight loss (an endpoint of neurodegeneration)
in male and female mice compared to untreated controls. Motor function was evaluated by performance on
a Rota-Rod. Tamoxifen maintained Rota-Rod performance for about an extra week. Vitamin E treatment signifi-
cantly delayed weight loss in females only. Rota-Rod performance was maintained slightly longer in mice treated
with vitamin E. Simultaneous use of both treatments did not delay weight loss longer than tamoxifen-only treat-
ment but had a greater effect than either treatment alone on Rota-Rod performance and demonstrated a significant
positive effect on the early "learning curve" portion of the Rota-Rod evaluations. We found significant but rela-
tively small improvements in rate of disease progression by treating Npc1-/- mice with tamoxifen and/or vitamin
E. Some sex differences in response and an early improvement in Rota-Rod performance suggest areas for further
study.
Key words: mice, neurodegeneration, Niemann-Pick C, Rota-Rod, tamoxifen, vitamin E.
protein was found to contain a sterol-sensingdomain consensus site and other motifs. This sug- Niemann-Pick disease type C (NPC) is gests a direct causative role for a mutant NPC1 a panethnic autosomal recessive disorder of un- product in the altered cholesterol movement in known pathogenesis (Vincent et al. 2003). A ma- NPC. In spite of this likely role, the patho- jor biochemical finding in this disorder is physiological basis for the symptoms present in the intracellular accumulation of unesterified cho- NPC is unknown and gangliosides accumulate lesterol within lysosomes and the Golgi apparatus.
as well (Zervas et al. 2001).
These findings have prompted the conclusion that Previous studies with npc1-/- mice revealed NPC is a disorder of intracellular cholesterol traf- a time-dependent accumulation of unesterified ficking (Patterson et al. 1995). The NPC1 gene has cholesterol in every organ except the brain. Subse- recently been cloned in man (Carstea et al. 1997) quently, however, it was found that the brain's ap- and mice (Loftus et al. 1997), and the predicted parent failure to accumulate cholesterol was due to Received: May 17, 2004. Accepted: July 6, 2004.
Correspondence: R.P. Erickson, Department of Pediatrics/4341B., 1501 N. Campbell Avenue, P. O. Box 245073, Tucson, Ari-zona 85724-5073, e-mail: [email protected] E.C. BascuZan-Castillo et al.
of heterozygous animals. Animals were kept at non-neuronal loss due to demyelination (Dietschy the University of Arizona Animal Care Facility and Turley 2001). Treatment of NPC patients with (PHS Assurance No. A-3248-01) on mouse chow agents that lower somatic cholesterol has not had containing 6% fat (or 10% for breeding mothers) significant effects on the neurological symptoms and water ad libitum. At weaning (at about 21 days (Patterson et al. 1993), although dimethyl of age), tail tips were removed from mice sulfoxide showed clinical improvement in one pa- and DNA was prepared. Polymerase chain reac- tient (Sakuragawa et al. 1988), and cholesty- tions (PCRs) to identify genotypes at the Npc1NIH ramine and lovastatin had short-term benefits as locus were performed using the primer pairs de- assessed by magnetic resonance imaging (Sylvain scribed in footnote 28 of Loftus et al. (1997).
et al. 1994). Nifedipine and probucol two agents For PCRs we used 10 mmol/L Tris, pH 8.3, that effectively reduce liver cholesterol did not al- 50 mmol/L KCl, 2.5 mmol/L Mg2+; 200 ìmol/L ter the progression of CNS disease in npc1-/- mice dNTPs, 1.25U Taq polymerase, and 1 ìmol (Erickson et al. 2000). It is not certain whether of each primer. DNA (20-40 ng) was added at most lipid-lowering drugs successfully permeate 85°C, and cycles of 30 s at 95°C, 30 s at 61°C, the blood-brain barrier, but a recent study found 1 min at 72°C ´ 35, and 10 min at 72°C were used.
The products were separated on 1.2% NuSieve terol-mobilizing cyclodextrins decreased liver agarose gels.
cholesterol storage in npc1-/- mice, but this route orintrathecal delivery had only slight effects on on- set of neurological symptoms (Camargo et al.
2001).
Two different drugs were investigated as a treat- In this study we analyzed the effects of ment in disease progression. Tamoxifen (desic- tamoxifen and vitamin E on the course of the dis- cate, 99% concentrate, from Sigma-Aldrich, St.
ease in Npc1-/- mice. Although tamoxifen is Louis, MO.) was prepared by dissolving 16 mg mostly known as an anti-estrogen used in the treat- of the drug per 1 ml of peanut oil at room tempera- ment and prevention of breast cancer, it has multi- ture (23oC). Following dissolution, tamoxifen was ple other effects relevant to NPC. Tamoxifen filtered with the 0.8/0.2 ìm pre-filter, filter combi- retards glycososphingolipid metabolism (Cabot nation (Gelman, Suporfilter #4905) and trans- et al. 1996) by inhibiting ceramide glycosylation ferred to a sterile container. Peanut oil used as (Lavie et al. 1997). It has lysosomotropic proper- a control was simply filtered with a 0.8/0.2 ìm fil- ties, altering vesicular transport recycling and se- ter and transferred to a sterile container. D-alpha cretory pathways (Altan et al. 1999), which could tocopherol (brand "Natural Liquid Vitamin E"), accentuate or ameliorate the NPC1 defect in cho- amphiphiles mimics the NPC1 cellular phenotypeand endogenous amphiphiles could be dislodged by tamoxifen (Lange and Steck 1998). It inhibitsglutamate-induced mitochondrial depolarization The Npc1-/- mice were weaned, genotyped and (Hoyt et al. 2000) and lowers serum cholesterol separated according to sex when they were ap- (Bilimoria et al. 1996). Orally administered proximately 21 days old. Treatment was initiated tamoxifen demonstrated a marked decrease in at this time. The mice were divided into four the development of lipid lesions in apolipoprotein groups: control, tamoxifen treated, vitamin E E knockout mice (Reckless et al. 1997). Finally, it treated, and tamoxifen plus vitamin E treated.
can reduce oxidative damage (Custodio et al.
1994). Vitamin E was chosen as another intraperitoneally with the drug suspended in pea- nut oil on a weekly basis with a dosage delivering0.023 ìg/g/day (0.01 c.c./g. and assuming linearrelease). Mice receiving vitamin E were given Material and methods
70 ìl (25 IU) of D-alpha tocopherol through oraladministration via a pipette on a weekly basis.
Control mice were either injected with plain pea- Npc1NIH mutant mice from the BALB/cJ back- nut oil (with a volume equal to tamoxifen treated ground were maintained by brother-sister mating mice) or were given no treatment.
Tamoxifen and vitamin E in NPC1 mice mal weight; < 10 seconds for 3 trials forRota-Rod), the last time when the mouse crossed All mice were evaluated using two criteria: weight the threshold was used as the age of failure, i.e.
loss and Rota-Rod performance. Mouse body a mouse might re-gain some weight or improve weights were recorded on a Monday, Wednesday, on the Rota-Rod for a measurement before again Friday schedule. When the weight of any mouse dropping below the threshold. In the case of dropped to 80% of its maximum value, rapid dete- a mouse's death before reaching the weight failure rioration and disease progression were observed.
point, the day of death was considered the failure This threshold was used as the weight failure point. The median age of failure was then used as point, when found in 2 consecutive measurements, an estimate of the survivability of the group.
for Kaplan–Meyer analysis, and mice were The chi-squares compare the Kaplan-Meyer prod- euthanized shortly thereafter.
uct for each group.
a Rota-Rod test instrument (Ugo Basile, NY,USA). The Rota-Rod was driven at a constant rate of 25 rpm. The mouse was given 3 trials onthe Rota-Rod, and the maximum trial time was re- Control mice did not show any variation in disease progression (according to the criteria measured), the Rota-Rod for 10 seconds were considered to comparing those treated with plain peanut oil ver- have failed.
sus those with no treatment. Consequently, no dis-tinction was made between the two groups.
Weights and Rota-Rod times for days on which measurements were not taken were calculated via A pilot study with tamoxifen injected subcutane- linear interpolation. The data were then analyzed ously showed a significant delay in the time by using Kaplan-Meyer survivability plots, as well of weight loss (data not shown). However, some as intra-group averaging for the Rota-Rod data.
leakage along the needle track led to questions The Kaplan-Meyer plots show the percentage about precise dosage. Hence the current study with of mice considered viable versus age. In the case intraperitoneal injections was performed.
of fluctuation around a failure point (80% of maxi- Table 1. Median weight failure points in days from Kaplan-Meyer survival plots
Vitamin E females Tamoxifen & vitamin E Table 2. Median Rota-Rod failure points in days from Kaplan-Meyer survival plots
Tamoxifen & vitamin E E.C. BascuZan-Castillo et al.
a significant difference in disease progression,beating their control counterparts by 8 days(6072 days, 95% CI, Table 1). However, the com-bined treatment was perhaps slightly less effica-cious than tamoxifen alone.
The median Rota-Rod failure results are presentedin Table 2. Control mice showed no significant sexdifferences (58 days, 5661 days 95% CI). Amongtamoxifen-treated mice, there was a slight delaybefore failure on the Rota-Rod (64 days,5171 days 95% CI). Mice treated with vitamin Eshowed a 7-day improvement over controls, witha median failure age of 65 days (62–66 days,95% CI). Mice treated with vitamin E were notsignificantly different from mice treated withtamoxifen. Mice given both treatments, showedan 11-day increase over controls with a medianfailure age of 69 days (5874 days, 95% CI; Ta-ble 2). In addition, these mice showed a significantincrease in absolute Rota-Rod performance (timespent on rotating rod) in the early stages of diseaseprogression (Figure 1).
It is now clear that the neurodegeneration seen in Figure 1. Average Rota-Rod performance for each group NPC is an autonomous process in the central ner- on an absolute scale: (A) male mice; (B) female mice vous system. The impact of visceral pathology onthe neurodegeneration in NPC was studied by Lof- The median weight failure points from tus et al. (2002). They reintroduced the wild-type Kaplan-Meyer plots are presented in Table 1.
NPC1 gene into npc1-/- mice by targeting its ex- The experimental and control groups did not show pression primarily to the CNS through the use of any sex differences (except with vitamin E) in dis- the prion protein promoter. Interestingly, neuro- ease progression and are grouped together.
degeneration was prevented, life span was normal- The control mice showed a median failure age of ized, and the sterility of npc1-/- mice was 59 days (57–61 days, 95% CI). Mice treated with corrected. The rescue did not completely rectify tamoxifen showed a significant improvement over the accumulation of GM2 or GM3 gangliosides in control mice. There was a 10-day improvement some neurons and glia (Loftus et al. 2002).
over controls with a median failure age of 69 days This observation is in agreement with other evi- (52–79 days, 95% CI; Table 1). Female mice dence that these higher order gangliosides are not treated with vitamin E showed the most significant obligatory players in NPC neurodegeneration (see sex difference in disease progression. Male mice below). Moreover, the persistence of visceral pa- showed no improvement when treated with vita- thology in the "rescued mice" reinforces the no- min E, having a median failure age of 57 days tion of an autonomous pathological process (56–58 days, 95% CI). By contrast, females occurring in the NPC brain. The cause of neuronal death is unclear. As a means of identifying the control mice by 11 days with a median failure the mode of neuronal death in NPC, Erickson and age of 70 days (67–74 days, 95% CI). In compari- Bernard (2002) overexpressed Bcl2, an anti- son to the tamoxifen group, the males in this group apoptotic protein in transgenic mice, using fared worse while the females fared about the neuronal specific enolase promoter. Bcl2 pre- the same. Mice given both treatments also showed vents developmental programmed cell death, and Tamoxifen and vitamin E in NPC1 mice neuronal death caused by a variety of stimuli.
A re-investigation of these mice showed that Cross breeding the neuronally-expressing Bcl2 unesterified cholesterol accumulation in the cere- mice with npc1-/- resulted in overexpression of Bcl2 in npc1-/- mouse brain neurons, but neuronal subcortical regions was markedly decreased and death was not spared (Erickson and Bernard that a few mice lived longer than controls 2002). When the mice were treated with (Gondre-Lewis et al. 2003). A pharmacological minocycline, a tetracycline analog that crosses approach was used to target another key synthetic the blood-brain barrier, and reduces neuronal enzyme higher upstream in the glycosphingolipid death in ischemia and Huntington and Parkinson synthetic pathway: glucosylceramide synthase.
diseases, neurodegeneration proceeded as in un- Oral administration of the inhibitor of this en- treated mice. This, and normal caspase1 levels in zyme, N-butyldeoxynojirimycin, to npc1-/- mice npc1-/- mouse brain, have suggested that neuronal and cats resulted in reduced ganglioside accumu- death in NPC does not proceed by a Bcl2 lation in the brain, accompanied by a modest delay and minocycline-inhibitable apoptotic pathway.
in onset of neurological dysfunction and death Thus, attempts to slow the progression of the animals, and reduced Purkinje cell loss of the disease with therapies targeted at other (Zervas et al. 2001).
pathophysiological mechanisms seem warranted.
We found that tamoxifen treatment signifi- As mentioned in the introduction, tamoxifen has cantly delayed the characteristic loss of weight a number of effects that could potentially amelio- that occurs in Npc1-/- mice (Table 1). This effect rate the neurodegeneration of NPC. The multiple was somewhat greater in males than in females properties of tamoxifen may be relevant to its use (which is not surprising given other sex differ- in the current "cocktail" now in clinical trials for ences in the symptoms of disease) (Erickson et al.
amyotrophic lateral sclerosis (Muscular Dystro- 2002) but not significantly so and our results are phy Association – http://www.azstarnet.com/star/ from both sexes pooled. Tamoxifen by itself had a smaller effect on Rota-Rod performance, which The mode of action of tamoxifen in ameliorat- assesses coordination versus ataxia. These results ing the symptoms of Npc1-/- is moot. The effects could suggest that the characteristic loss of weight of tamoxifen on ceramide glycosylation (Cabot is not merely due to an inability to feed from et al. 1996, Lavie et al. 1997) could still be rele- the overhead water and pellet containers. In con- vant to NPC. There is substantial accumulation of trast to our finding of a slight benefit of GM2, GM3, and other glycosphingolipids in the intraperitoneal tamoxifen on neurodegeneration in NPC brain, which has prompted studies of the role Npc1-/- mice, intracranial administration of of these lipids in NPC neuropathogenesis.
tamoxifen to rat pups from 6 to 9 days of age inhib- Taniguchi et al. (2001) found that GM1 accumu- ited Purkinje dendritic outgrowth (Sakamoto et al.
lates primarily in neurons and astrocytes of npc1-/- 2003). Thus, of these 2 drugs, vitamin E might be mouse brain, whereas GM2 accumulates in neu- more appropriate for further study.
rons and macrophages. GM1, which is normally Vitamin E has long been known as an antioxi- localized to synaptosomal membranes of neurons, dant. Resent research has focused on particular accumulates in the cytoplasm and other dendrites, pathways that are affected by it, e.g. signaling whereas GM2, which is typically absent from the pathways (Rimbach et al. 2002) or superoxide pro- neuronal soma, accumulates in perinuclear vesi- duction (Ulker et al. 2003). We found no effect of cles. In view of the postulated role of these vitamin E in male mice, while female mice showed a significant delay in weight loss. How- dendritogenesis, these subcellular alterations ever, sex differences were insignificant for alone may contribute to neuronal dysfunction.
Rota-Rod performance so that data could be A genetic approach aimed at reducing the lev- pooled and vitamin E treated mice showed a mild els of gangliosides by mating npc1-/- mice with improvement in this parameter.
Effects of tamoxifen plus vitamin E were gen- the â-1-4GalNAc transferase gene responsible for erally similar to those of tamoxifen or vitamin E synthesis of GM2 and higher order gangliosides, for both growth and motor performance. The com- successfully reduced CNS accumulation of GM2 bined treatment was slightly better than either and glycolipids GA1 and GA2, but did not im- alone for Rota-Rod performance but not better for prove the clinical phenotype or neuronal pathol- weight maintenance. Perhaps it is the combined ogy of the npc1-/- mice (Liu et al. 2000).
E.C. BascuZan-Castillo et al.
anti-oxidant influence of these 2 drugs which is Custódio JB, Dinis TC, Almeida LM, Madeira VM, intramembranous inhibitors of lipid peroxidation.
When analyzing Rota-Rod average perfor- Evidence for peroxyl radical scavenging activity.
mance (Figure 1) instead of Rota-Rod "time of Biochem Pharmacol 47: 1989–1998.
failure" (Table 2), an enhancing effect of the com- Dietschy JM, Turley SD, 2001. Cholesterol metabolism bination of tamoxifen and vitamin E on the early in the brain. Curr Opin Lipidol 12: 105–112.
"learning" part of the curve was seen. Using Erickson RP, Bernard O, 2002. Studies on neuronal Rota-Rod in the constant speed mode, normal death in the mouse model of Niemann-Pick C dis- mice improve their performance in the first ses- ease. J Neurosci Res 68: 738–744.
sions as they learn to walk on the rotating rod but Erickson RP, Garver WS, Camargo F, Hossain GS, not on the accelerating Rota-Rod (Jones et al.
Heidenreich RA, 2000. Pharmacological and ge-netic modifications of somatic cholesterol do not 1968). Only Npc1-/- treated with tamoxifen and vi- substantially alter the course of CNS disease in tamin E show such an early learning component.
Niemann-Pick C mice. J Inher Metab Dis 23: This effect suggests that this drug combination may be delaying the dementia that is seen in hu- Erickson RP, Kiela M, Devine PJ, Hoyer PB, mans but has been little studied in mice (Voikar et Heidenreich RA, 2002. mdr1a deficiency corrects sterility in Niemann-Pick C1 protein deficient fe-male mice. Mol Reprod Dev 62: 167–173.
Gondre-Lewis MC, McGlynn R, Walkley SU, 2003.
Cholesterol accumulation in NPC1-deficient neu- rons is ganglioside dependent. Curr Biol 13:1324–1329.
We have found significant but relatively small im- provements in rate of disease progression by treat- Reynolds IJ, 2000. Inhibition of glutamate-induced ing Npc1-/- mice with tamoxifen and/or vitamin E.
mitochondrial depolarization by tamoxifen in cul-tured neurons. J Pharmacol Exp Ther 293: 480–486.
The mechanism may be due to the combined Jones BJ, Roberts DJ, 1968. The quantitative measure- anti-oxidant effect of these 2 agents. Some sex dif- ment of motor inco-ordination in naïve mice using ferences in response and an early improvement in an accelerating rotarod. J Pharm Pharmac 20: Rota-Rod performance suggest areas for further Lange Y, Steck TL, 1998. Four cholesterol-sensing proteins. Curr Opin Struct Biol 8: 435–439.
Acknowledgements. This research was funded by
Lavie Y, Cao H, Volner A, Lucci A, Han TY, Geffen V, NIH grant R01-EB00343 and the Ara Parseghian et al. 1997. Agents that reverse multidrug resis- Medical Research Foundation. We thank Donelle tance, Tamoxifen, Verapamil, and Cyclosporin A, Myers for secretarial support and Kathylynn Saboda block glycosphingolipid metabolism by inhibiting for help in statistical analyses.
ceramide glycosylation in human cancer cells.
J Biol Chem 272: 1682–1687.
Liu Y, Wu YP, Wada R, Neufeld EB, Mullin KA, Howard AC, et al. 2000. Alleviation of neuronalganglioside storage does not improve the clinical Altan N, Chen Y, Schindler M, Simon SM, 1999.
course of the Niemann-Pick C disease mouse. Hum Tamoxifen inhibits acidification in cells independ- Mol Genet 9: 1087–1092.
ent of the estrogen receptor. Proc Nat Acad Sci, Loftus SK, Erickson RP, Walkley SU, Bryant MA, USA 96: 4432–4437.
Incao A, Heidenreich RA, Pavan WJ, 2002. Rescue Bilimoria MM, Assikis VJ, Jordan VC, 1996. Should of neurodegeneration in Niemann-Pick C mice by adjuvant tamoxifen therapy be stopped at 5 years? a prion-promoter-driven Npc1 DNA transgene.
Cancer J Sci Am 2: 140–150.
Hum Mol Genet 11: 3107–3114.
Cabot MC, Giuliano AE, Volner A, Han TY, 1996.
Tamoxifen retards glycosphingolipid metabolism Cummings C, Brown A, et al. 1997. Murine model in human cancer cells. FEBS Lett 394: 129–131.
of Niemann-Pick C disease: mutation in a choles- Camargo F, Erickson RP, Garver WS, Hossain GS, terol homeostasis gene. Science 277: 232–235.
Carbone PN, Heidenreich RA, Blanchard J, 2001.
Patterson MC, Di Bisceglie AM, Higgins JJ, Abel RB, Cyclodextrins in the treatment of a mouse model Schiffmann R, Parker CC, et al. 1993. The effectof of Niemann-Pick C disease. Life Sci 70: 131–142.
and plasma cholesterol in Niemann-Pick disease Carstea ED, Morris JA, Coleman KG, Loftus SK, type C. Neurology 43: 61–64.
Zhang D, Cummings C, et al. 1997. Niemann-Pick Patterson, MC, Pentchev PG, Venier MT, Suzuki K, C1 disease gene: homology to mediators of choles- 1995. Niemann-Pick disease type C: a cellular cho- terol homeostasis. Science 277: 228-231.
lesterol lipidosis. In: Scriver C, Beaudet AL, Tamoxifen and vitamin E in NPC1 mice Sly WS, Valle D, eds. The metabolic and molecular cholestyramine and lovastatin. Pediatr Neurol 10: bases of inherited disease, Vol. 2, 7th edn McGraw-Hill: New York: 2625–2639.
Taniguchi M, Shinoda Y, Ninomiya H, Vanier MT, Reckless J, Metcalfe JC, Grainger DJ, 1997. Tamoxifen Ohno K, 2001. Sites and temporal changes of decreases cholesterol sevenfold and abolishes lipid lesion development in apolipoprotein E knockout Niemann-Pick disease type C mouse brain. Brain mice. Circulation 95: 1542–1548.
Dev 23: 414–421.
Rimbach G, Minihane AM, Majewicz J, Fischer A, Ulker S, McKeown PP, Bayraktutan U, 2003. Vitamins Pallauf J, Virgli F, Weinberg PD, 2002. Regulation reverse endothelial dysfunction through regulation of cell signalling by vitamin E. Proc Natl Acad Sci of eNOS and NAD(P)H oxidase activities. Hyper- USA 61: 415–425.
tension 41: 534–539.
Sakamoto H, Mezaki Y, Shikimi H, Ukena K, Vincent I, Bu B, Erickson RP, 2003. Understanding Tsutsui K, 2003. Dendritic growth and spine forma- Niemann-Pick type C disease: a fat problem. Curr tion in response to estrogen in the developing Opin Neurol 16: 155–161.
Purkinje cell. Endocrinology 144: 4466–4477.
Voikar V, Rauvala H, Ikonen E, 2002. Cognitive deficit Sakuragawa N, Ohmura K, Suzuki K, Miyazato T, and development of motor impairment in a mouse 1988. Clinical improvement with DMSO treatment model of Niemann-Pick type C disease. Behav in a patient with Niemann-Pick disease (type C).
Brain Res 132: 1–10.
Acta Paediatr Japan 30: 509–516.
Zervas M, Somers KL, Thrall MA, Walkley SU, 2001.
Sylvain M, Arnold DL, Scriver CR, Schreiber R, Shevell MI, 1994. Magnetic resonance spectros- Niemann-Pick disease type C. Curr Biol 11: copy in Niemann-Pick disease type C: correlation

Source: http://jag.igr.poznan.pl/2004-Volume-45/4/pdf/2004_Volume_45_4-461-467.pdf