Gene 536 (2014) 97–104

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r .com/ locate /gene

Reduced glucosylceramide in the mouse model of Fabry disease:Correction by successful enzyme replacement therapy

Rui Quinta a,b,⁎, Daniel Rodrigues a, Marisa Assunção a, Maria Fatima Macedo a,c, Olga Azevedo b,d,Damião Cunha b, Pedro Oliveira e, Maria Clara Sá Miranda a

a Lysosome and Peroxisome Biology Unit (UniLiPe), IBMC - Instituto de Biologia Molecular e Celular, Universidade do Porto, Rua do Campo Alegre no. 823, 4150-180, Porto, Portugalb School of Health Sciences, University of Minho, Campus de Gualtar, 4710-057, Braga, Portugalc SACS, University of Aveiro, Aveiro, Portugald Cardiology Department, Centro Hospitalar do Alto Ave, Rua dos Cutileiros, 4835-044 Guimarães, Portugale Instituto de Ciências Biomédicas Abel Salazar, Universidade do Porto, Rua de Jorge Viterbo Ferreira no. 228, 4050-313 Porto, Portugal

Abbreviations: GlucCer, glucosylceramide; Cer, ceramtherapy; Gb3, globotriaosylceramide; GSL, glycosphingoliease; Lyso-Gb3, globotriaosylsphingosine.⁎ Corresponding author at: Lysosome and Peroxisome

Instituto de Biologia Molecular e Celular, Rua do CampPortugal. Tel.: +351 226074900; fax: +351 226099157.

E-mail addresses: ruiquinta@ibmc.up.pt (R. Quinta), d(D. Rodrigues), marisa.assuncao@ibmc.up.pt (M. Assunçã(M.F. Macedo), olgazevedo@yahoo.com.br (O. Azevedo), d(D. Cunha), pnoliveira@icbas.up.pt (P. Oliveira), mcsamir@

0378-1119/$ – see front matter © 2013 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.gene.2013.11.073

a b s t r a c t

a r t i c l e i n f o

Article history:Accepted 26 November 2013Available online 13 December 2013

Keywords:Fabry diseaseEnzyme replacement therapyCeramideGlucosylceramideFabry mice

Fabry disease is an X-linked lysosomal storage disease (LSD) caused by deficient activity of α-GalactosidaseA (α-Gal A). As a result, glycosphingolipids, mainly globotriaosylceramide (Gb3), progressively accumulatein body fluids and tissues. Studies aiming at the identification of secondary lipid alterations in Fabry disease maybe potentially useful for the monitorization of the response to enzyme replacement therapy (ERT) and develop-ment of future therapies. The focus of this study was to evaluate if α-Gal A deficiency has an effect on two keygroups of molecules of sphingolipids metabolism: glucosylceramides (GlucCers) and ceramides (Cers). Studiesperformed in amousemodel of Fabry disease showed reduced level of GlucCer and normal level of Cer in plasma,liver, spleen, kidney and heart. Moreover, analysis of GlucCer isoforms in Fabry knockout mice showed thatGlucCer isoforms are unequally reduced in different tissues of these animals. ERT had a specific effect on theliver's GlucCer levels of Fabry knockout mice, increasing hepatic GlucCer to the levels observed in wild typemice. In contrast to Fabry knockout mice, plasma of Fabry patients had normal GlucCer and Cer but an increasedGlucCer/Cer ratio. This alteration showed a positive correlationwith plasmaglobotriaosylsphingosine (lyso-Gb3)concentration. In conclusion, this work reveals novel secondary lipid imbalances caused by α-Gal A deficiency.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Fabry disease is an X-linked lysosomal storage disease (LSD)caused bymutations in the GLA gene encoding the lysosomal hydrolaseα-Galactosidase A (α-Gal A). Reported incidences, ranging from1:476,000 (Poorthuis et al., 1999) to 1:117,000 (Meikle et al., 1999) inthe general population, may largely underestimate the true prevalenceof the disease. Indeed, recent estimated prevalences in Fabry newbornscreenings range from 1:3100 (Spada et al., 2006) to 1:13,341(Wittmann et al., 2012), suggesting that Fabry disease is more frequentthan previously anticipated. Loss-of-function ofα-Gal A results in a pro-gressive accumulation of substrates with terminal α-galactosyl

ide; ERT, Enzyme replacementpid; LSD, lysosomal storage dis-

Biology Unit (UniLiPe), IBMC -o Alegre, 823, 4150-180 Porto,

anielr@ibmc.up.pto), fmacedo@ibmc.up.ptamiaojlcunha@gmail.comibmc.up.pt (M.C. Sá Miranda).

ights reserved.

moieties, particularly globotriaosylceramide (Gb3), in bodily fluidsand in the lysosomes. The accumulation of lipid substrates causes amultisystemic disease with a wide array of clinical signs and symptomsthat predominantly affect the heart, central nervous system and kidney.Despite being an X-linked disorder, heterozygous female patients alsooften display clinical symptoms (Wang et al., 2007). With age, progres-sive damage to major organ systems occurs, culminating in severe car-diac and cerebrovascular complications, end-stage renal failure andpremature mortality (Germain, 2010).

Enzyme replacement therapy (ERT)with recombinantα-Gal A is theprimary treatment for Fabry disease. It has been shown that ERT atten-uates renal and cardiac clinical manifestations of Fabry patients(Banikazemi et al., 2007). Moreover, ERT reduces the primary accumu-lation of Gb3 (Schiffmann et al., 2006; Young et al., 2005) that was, for along time, considered as a surrogate marker of Fabry disease. Indeed,elevated levels of Gb3 are frequently present in urine of patientswith Fabry disease (Fuller et al., 2005) and plasma of classic Fabryhemizygotes (Togawa et al., 2010; Young et al., 2005). However, plasmaGb3 is variably elevated inmaleswithmilder variant forms of Fabry dis-ease and heterozygotes (Togawa et al., 2010; Young et al., 2005). Fur-thermore, several studies also indicate that Gb3 levels poorly reflectFabry disease manifestations and therapeutic outcome (Bekri et al.,

98 R. Quinta et al. / Gene 536 (2014) 97–104

2006; Schiffmann et al., 2013;Whitfield et al., 2005; Young et al., 2005).Recently, highly elevated levels of globotriaosylsphingosine (lyso-Gb3)were discovered in plasma (Aerts et al., 2008) and urine (Auray-Blaiset al., 2010) of patients with Fabry disease. Plasma lyso-Gb3 levels cor-relatewith the risk of cerebrovascular disease in hemizygotes and over-all disease severity of heterozygotes (Rombach et al., 2010), suggestingthat it may play a direct role in the pathogenesis of Fabry disease.

Apart from primary lipid accumulation, considerable evidence sug-gests that secondary alterations in lipid compounds, without a directlink to the primary protein defect, can themselves be actively involvedin the pathogenesis of several LSDs (Walkley and Vanier, 2009). More-over, the identification of secondary lipid alterations in LSDsmay poten-tially contribute to the discovery of novel ways of discerning patientsfrom healthy subjects and monitoring therapeutic efficacy. Indeed, pre-vious studies evaluating the efficacy of therapy on animal models ofMucopolysaccharidoses (MPS) (Ciron et al., 2006; Cressant et al.,2004; Desmaris et al., 2004; Ellinwood et al., 2007; Frisella et al.,2001) andNiemann–Pick type C (NPC) (Griffin et al., 2004) demonstrat-ed correlations between the normalization of secondary lipid alterationsand the amelioration of primary biochemical markers or clinical fea-tures. To our knowledge, similar studies have not been performed inthe mouse model of Fabry disease (α-Gal A knockout mice), which isa valuable tool to elucidate underlying mechanisms of Fabry diseasepathophysiology. Despite having a normal lifespan without organ fail-ure (Ohshima et al., 1999), biochemically, Fabry knockout mice havebeen shown to reproduce abnormalities found in Fabry patients, includ-ing a significantGb3 (Rodrigues et al., 2009) and lyso-Gb3 accumulation(Aerts et al., 2008). Therefore, it is of considerable interest to identifysecondary lipid alterations associated with the mouse model of Fabrydisease.

In patients with Fabry disease, secondary urinary alterations in Cer-amide (Cer) and Glucosylceramide (GlucCer) isoforms, with specificacyl-chain lengths, have been reported (Fuller et al., 2005). More specif-ically, secondary increases in Cer 16:0 and 24:1 in Fabry hemizygotesand decreases in GlucCer 22:0 and 24:0 in heterozygotes were found.Apart from its role as a central molecule in glycosphingolipid (GSL)me-tabolism (Fig. 1), Cer is a bioactive molecule and a mediator in cellularsenescence, differentiation, apoptosis (Hannun and Obeid, 2008) andinflammation (Wu et al., 2007). In contrast to the growth-inhibitory ef-fects of Cer, GlucCer (Fig. 1) exerts pro-mitogenic effects in a variety oftissues and cell types (Datta and Radin, 1988; Marsh et al., 1995;Shayman et al., 1991). Therefore, these bioactive molecules appear to

Fig. 1. Biosynthesis flow chart of GSLs from Cer. Impairedα-Galactosidase A (α-Gal A) activity leand subsequent cellular dysfunction. Also depicted are the enzymes responsible for the synthesisynthase; LCS—Lactosylceramide synthase; Gb3S—Globotriaosylceramide synthase.

be critical for the maintenance of cell and tissue homeostasis. Thepresent studywas designed to evaluate ifα-Gal A deficiency causes sec-ondary alterations in GlucCer and Cer, including their respective iso-forms, in the Fabry disease mouse model and in the plasma of patientswith Fabry disease.

2. Materials and methods

2.1. Animal studies

α-Gal A knockout mice were provided by the National Institutes ofHealth (Bethesda, MD, USA) and a colony was maintained at theInstituto de Biologia Molecular e Celular (IBMC, Porto, Portugal). Giventhe genetic background of this mouse colony, C57BL/6J mice wereused as controls. Food and water were provided ad libitum. The IBMCanimal facility is licensed for breeding and experiments with laboratoryrodents. Animal experiments were carried out in accordance with theEU Directive 2010/63/EU. All animals used in this study were 12-week-old males.

For the ERT experiments, α-Gal A (agalsidase alfa, Replagal©, ShireHuman Genetics) was administered weekly to Fabry knockout miceby intravenous injection, via the tail vein, of 100 μL at 1.5 or 0.2 mg/kgbody weight, from the fourth to the eleventh week of life. 8 weekly ad-ministrations of 1.5 mg/kg α-Gal A were selected because it has previ-ously been demonstrated that this protocol successfully reduces Gb3accumulation in Fabry knockout mice (Macedo et al., 2012). The dosageof 0.2 mg/kg was chosen because it is the approved clinical dosage ofagalsidase alfa. One group of Fabry knockoutmice receivedweekly infu-sions of saline solution during the same period. At week twelve, animalswere sacrificed, organs were harvested and kept at −80 °C until use.Sample sizes were determined based on a power analysis. Power analy-sis was performed taking into consideration previous information re-garding the standard deviations (our own unpublished data), for asignificance level of 0.05 and power of 0.8. The total numbers of animalsused in this study were: 11 wild type mice; 7 Fabry knockout mice; 5Fabry knockout mice treated with 0.2 mg/kg α-Gal A; and 3 Fabryknockout mice treated with 1.5 mg/kg α-Gal A.

2.2. Patient samples

Plasma samples fromunaffected individuals and Fabry patientswerekept at −20 °C until analysis. The control groups contained 20 males

ads to the primary accumulation of Globotriaosylceramide (Gb3, red box) in Fabry diseases of GlucCer from Cer, LacCer fromGlucCer, and Gb3 from LacCer: GCS—Glucosylceramide

Table 1Cer and GlucCer transitions monitored in MRM mode.

Analyte m/z Values Cone (V) Collision (v)

Cer 12:0 (IS) 482.41 N 264.4 25 15Cer 14:0 510.70 N 264.4 25 15Cer 16:0 538.70 N 264.4 27 20Cer 18:1 564.50 N 264.4 27 13Cer 18:0 566.40 N 264.4 27 13Cer 20:0 594.70 N 264.4 27 15Cer 22:0 622.80 N 264.4 27 15Cer 24:1 650.00 N 264.4 27 15Cer 24:0 651.60 N 264.4 27 15GlucCer 12:0 (IS) 644.22 N 264.4 30 25GlucCer 16:0 700.70 N 264.4 30 25GlucCer 18:0 728.80 N 264.4 30 26GlucCer 20:0 756.90 N 264.4 30 26GlucCer 21:0 770.80 N 264.4 30 26GlucCer 22:0 784.90 N 264.4 30 27GlucCer 23:0 798.80 N 264.4 30 27GlucCer 24:1 810.90 N 264.4 30 27GlucCer 24:0 812.90 N 264.4 30 27

99R. Quinta et al. / Gene 536 (2014) 97–104

and 12 females. The diagnosis of Fabry disease was confirmed by re-duced α-Gal A activity in dried blood spots, leukocytes, plasma and/orgenotyping. Blood samples from 9 hemizygous and 8 heterozygousnaïve (not on ERT) Fabry patients were obtainedwith informed consentand in accordance with the Code of Ethics of theWorld Medical Associ-ation (Declaration of Helsinki).

2.3. Quantification of GlucCer and Cer by ultra-performance liquid chroma-tography coupled to tandem mass spectrometry (UPLC–MS/MS)

2.3.1. GlucCer and Cer extractionSphingolipids from plasma and tissue homogenates were extracted

according to previously described protocols (Bielawski et al., 2006;Hammad et al., 2010), withminormodifications. Non-natural occurringinternal standards (IS) and endogenous sphingolipid calibrator stan-dards for UPLC–MS/MS experiments were obtained from Avanti PolarLipids (Alabaster, AL, USA).

In brief, tissue homogenates and plasma samples were fortifiedwithamixture of IS Cer 12:0 and GlucCer 12:0 at a concentration of 200 ppb.Tissue homogenates were extracted with 2 mL of isopropanol/water/ethyl acetate (30:10:60, v/v/v), vortexed, sonicated thrice for 30 s, andcentrifuged for 10 min at 4000 rpm. The supernatant was transferredto a new vial and samples were re-extracted. Upper layers were com-bined and evaporated to dryness under N2 flow. For plasma analyses,100 μL of plasma were diluted to 2 mL with serum free media and ex-tracted with 2 mL of isopropanol/ethyl acetate (15:85, v/v). Sampleswere centrifuged for 10 min at 4000 rpm, the upper phase was trans-ferred to a new vial and acidified with 100 μL formic acid. Plasma sam-ples were re-extracted, the upper organic phases were combined andevaporated to dryness under N2.

2.3.2. Ultra-performance liquid chromatography coupled to tandem massspectrometry (UPLC–MS/MS)

Analyses of GlucCer and Cer of different acyl moieties wereperformed on a Waters (Milford, MA, USA) Acquity Ultra PerformanceLiquid Chromatography® (UPLC) system coupled to a Quattro PremierXE (Micromass® MS technologies). The sphingolipid extract wasreconstituted in methanol/1% formic acid/0.5% ammonium formate1 M (v/v/v) and 10 μL were injected in full loop mode into the system.Isoforms were separated on an Acquity® UPLC BEH C18 1.7 μm column(2.1 × 150 mm) with a methanol/water/1% formic acid/0.5% ammoni-um formate 1 M (v/v/v/v) gradient. Analyses were performed byelectrospray ionization tandem mass spectrometry in the positive ionmode to select both parent and daughter ions specific to each analyte,simultaneously, from a single injection. The transitions monitored inmultiple-reaction monitoring mode (MRM) are listed in Table 1. Theconfigurations for mass spectrometry were: capillary voltage, 3.5 kV;cone voltage, 30 V; source temperature, 120 °C; desolvation tempera-ture, 300 °C. All samples were run in duplicate and two reference sam-ples were included in every run.

Reference standards Cer 18:1 andGlucCer 18:0 (concentration range0–500 ppb) were added to matrix samples containing the respective IS.Linear calibration curves were generated by plotting the ratio of thepeak area of the reference standard to that of the IS versus increasingconcentrations of the respective reference standard.

The concentration of each isoform monitored was determined bylinear regression with the IS. The ratio of the peak area sum of the tran-sitions monitored to the area of the peak of the respective IS was alsodetermined and used to calculate the total GlucCer and total Cerconcentrations.

Because of the current unavailability of lipid standards to performabsolute quantifications for all isoforms monitored by MS, we normal-ized all our data regarding GlucCer and Cer isoforms to the mean valuesobtained with wild type mice (in animal experiments) and gender-matched healthy subjects (in human studies). This normalization

method has been used in other studies (Durant et al., 2011; Mullenet al., 2011; Schwab et al., 2008).

2.4. Gb3 quantification in Fabry knockout mice tissues

Gb3 was quantified by UPLC–MS/MS in Fabry knockout mice tissuesamples as previously described (Macedo et al., 2012).

2.5. Lyso-Gb3 quantification in Fabry tissues and plasma

Lyso-Gb3 extraction was performed according to a previously de-scribed method (Aerts et al., 2008), with minor modifications.Lactosylsphingosine (LacSph, Avanti Polar Lipids, Alabaster, AL, USA)was added as an IS to each sample at a concentration of 50 ppb. Totallipids were extracted with 2 mL of chloroform: methanol 2:1 (v/v)with thorough mixing. The extract was centrifuged at 14,000 ×g for10 min and the supernatant was removed. To the supernatant, water(400 μL) was added, samples were thoroughly vortexed and the twolayers were separated by centrifugation. The upper phasewas collected,dried under N2 flow and taken up with 1 mL of water. The water phasewas extracted twice with 1 mL of water-saturated 1-butanol and driedonce more under N2. Samples were reconstituted in methanol/1%formic acid/0.5% ammonium formate 1 M (v/v/v) and 10 μL wereinjected into the UPLC–MS/MS system in full loop mode. Separationwas performed on an Acquity® UPLC BEH C18 1.7 μm column(2.1 × 150 mm) with a methanol/water/1% formic acid/0.5% ammoni-um formate 1 M (v/v/v/v) gradient. The following transitions (m/z)were monitored in MRM mode: Lyso-Gb3, 786.40 N 282.72; LacSph,624.00 N 282.60. Calibration curves were generated by standard addi-tion of lyso-Gb3 (Avanti Polar Lipids, Alabaster, AL, USA) tomatrix sam-ples. The ratio of the peak area of lyso-Gb3 to the area of the peak ofLacSph was determined in each sample and used to calculate lyso-Gb3concentration based on the calibration curves. Results were normalizedto the amount of tissue/plasma in each sample.

2.6. Statistical analysis

Continuous variables were compared by independent Student'st-test and Mann–Whitney U test depending on whether normalityand/or homogeneity of variances were observed, assessed by theShapiro–Wilk and Levene tests. Comparison of multiple groups wasperformed using one-way ANOVA; pairwise comparisons were basedon Tukey's Multiple Comparison test. The statistical association of thevariation in total GlucCer/total Cer ratio with lyso-Gb3 concentrationin Fabry patients' plasmawas assessed by Spearman rank's order corre-lation. P b 0.05 was considered significant.

Table 2GlucCer isoforms in Fabry knockout comparatively to WT mice as analyzed by UPLC–MS/MS.

Fabry/WTbaseline

Plasma Spleen Liver Kidney Heart

GlucCer 16:0 0.3 ± 0.1** 0.4 ± 0.1*** 0.6 ± 0.3 0.7 ± 0.2* 0.3 ± 0.0***GlucCer 18:0 0.8 ± 0.1 0.3 ± 0.4* 0.9 ± 0.1 nq 0.3 ± 0.1***GlucCer 20:0 0.6 ± 0.1* 0.5 ± 0.1*** 0.4 ± 0.1* 0.7 ± 0.2* 0.6 ± 0.1***GlucCer 21:0 0.7 ± 0.1 nq nq nq 0.3 ± 0.2*GlucCer 22:0 0.6 ± 0.0 0.5 ± 0.1*** 0.6 ± 0.2** 0.7 ± 0.2* 0.4 ± 0.1***GlucCer 23:0 0.7 ± 0.0* 0.5 ± 0.1** 0.6 ± 0.1* 0.4 ± 0.1*** 0.4 ± 0.1***GlucCer 24:0 0.5 ± 0.1** 0.4 ± 0.1** 0.6 ± 0.2* 0.5 ± 0.2*** 0.4 ± 0.2***GlucCer 24:1 0.4 ± 0.0*** 0.4 ± 0.1*** 0.4 ± 0.1* 0.7 ± 0.2 0.4 ± 0.1***

Data are expressed as mean ± SD. For simplification purposes, values of WT mice foldchange relative to the WT baseline (1.0 ± SD for all monitored isoforms) were omitted.nq — below quantification limit. Statistically significant differences between WT andFabry knockout mice are indicated by *p b 0.05, **p b 0.01 and ***p b 0.001. Plasma:n = 5 animals per group; spleen: n = 11 WT and 7 Fabry; liver: n = 10 WT and 7Fabry; kidney: n = 9 WT and 7 Fabry; heart: n = 9 WT and 5 Fabry.

100 R. Quinta et al. / Gene 536 (2014) 97–104

Data were statistically analyzed using PASW® Statistics 18 software(IBM, Armonk, NY, USA).

3. Results

3.1. α-Gal A deficient mice present a decrease in GlucCer without alterationin Cer level

In order to assess the existence of secondary alterations in GlucCerand Cer in the Fabry disease mouse model, we evaluated the totallevel of these classes of molecules in different organs of wild type andFabry knockout mice by UPLC–MS/MS (Fig. 2).

Quantification of total GlucCer concentration revealed significant re-ductions in splenic (≈53%), hepatic (≈35%), renal (≈32%), cardiac(≈46%) (Fig. 2A) and plasma (≈44%) (Fig. 2B) GlucCer levels in Fabryknockoutmice comparatively to those of age-matched controls. Quanti-fication of total Cer in the same organs (Fig. 2C) and plasma (Fig. 2D) didnot show differences between the two groups of animals.

3.2. GlucCer isoform alterations in Fabry knockout mice

Taking into account the preceding results, we decided to analyze ifthe observed decrease in total GlucCer concentration of Fabry knockoutmice was the result of lower levels of a specific GlucCer isoform. To thisend, GlucCer isoformswere compared in plasma, spleen, liver, heart andkidney samples of wild type and Fabry knockout mice by UPLC–MS/MS(Table 2).

Plasma levels of GlucCer 16:0, 20:0, 23:0, 24:0 and 24:1 were re-duced in Fabry knockout mice comparatively to those of wild type ani-mals (Table 2). Pairwise comparisons between the different GlucCerisoforms revealed that plasma GlucCer 16:0 was significantly more re-duced than the other isoforms monitored except 24:1 (SupplementalTable A). In sharp contrast to what was observed in plasma, hepatic

Fig. 2. Total GlucCer and total Cer quantification inwild type (WT) and Fabry knockoutmice byWT and Fabry knockout mice. (C, D) Total Cer quantification in different organs (C) and pla**p b 0.01 and ***p b 0.001. Spleen: n = 11 WT and 7 Fabry; liver: n = 10 WT and 7 Fabry; kgroup.

levels of GlucCer 16:0 in Fabry knockout mice weren't significantly dif-ferent from controls. In this case, GlucCer isoforms with an acyl chainwith more than 18 carbons were lower comparatively to wild typemice (Table 2). The decreases in these isoformsweren't significantly dif-ferent fromeachother (Supplemental Table A). In the spleen andkidneyof Fabry knockout mice, all monitored isoforms, with the exception ofrenal GlucCer 24:1, were lower comparatively to those of controls(Table 2). Pairwise comparisons showed that all monitored isoformshad similar decreases in both organs (Supplemental Table A). In cardiacsamples of Fabry knockout mice, all GlucCer isoforms were found to besignificantly reduced (Table 2) and significant differences between thedecreases in the 16:0, 18:0, 21:0 isoforms and the 20:0 isoformwere ob-served (Supplemental Table A).

Overall, these results indicate that GlucCer isoforms are unequallyreduced in different tissues of Fabry knockout mice.

UPLC–MS/MS. (A, B) Total GlucCer quantification in different organs (A) and plasma (B) ofsma (D) of WT and Fabry knockout mice. Data are expressed as mean ± SD. *p b 0.05,idney: n = 9 WT and 7 Fabry; heart: n = 9 WT and 5 Fabry; plasma: n = 5 animals per

Fig. 3. Effect of ERT with α-Gal A in total GlucCer of Fabry knockout mice as analyzed byUPLC–MS/MS. Data are expressed as mean ± SD. **p b 0.01. n = 5 untreated Fabry, ex-cept for liver and kidney (n = 7). n = 5 Fabry + 0.2 mg/kg α-Gal A; n = 3Fabry + 1.5 mg/kg α-Gal A.

101R. Quinta et al. / Gene 536 (2014) 97–104

UPLC–MS/MS analyses of plasma, spleen, liver, kidney and heart Cerisoforms did not reveal significant differences between Fabry knockoutand wild type mice (data not shown).

3.3. Enzyme replacement therapy normalizes GlucCer levels in the liver ofFabry knockout mice

In order to investigate if ERT was able to correct the secondary de-crease in GlucCer observed in Fabry knockout mice, two groups ofFabry animals were injected weekly, from the fourth to the eleventhweek of life, with 0.2 mg/kg or 1.5 mg/kg α-Gal A. A control group ofFabry knockout mice was injected with saline solution during thesame period and the animals were sacrificed oneweek after the last en-zymatic administration.

Fabry disease biomarkers Gb3 and lyso-Gb3 were analyzed byUPLC–MS/MS to confirm that the ERT protocol used was effective incorrecting primary storage of α-Gal A substrates (Table 3).

As expected, treatment with α-Gal A resulted in an organ and dose-dependent reduction of Gb3 accumulation. Treatment with 0.2 mg/kgα-Gal A was sufficient to reduce Gb3 by 97% in the liver, 88% in thespleen, 69% in the kidney and 85% in the heart. Treatment with1.5 mg/kg α-Gal A further reduced Gb3 accumulation comparativelyto low-dose treated Fabry animals, with the exception of splenic Gb3,where high-dose treated animals showed a Gb3 level equivalent tothat of mice treated with the lower dose of α-Gal A. Administration of0.2 mg/kg α-Gal A markedly reduced lyso-Gb3 in all organs with theexception of the spleen (5% lyso-Gb3 reduction). Greater reductions oftissue lyso-Gb3 levels were seen in the group of mice treated with1.5 mg/kg α-Gal A (below the detection limit in all organs studied,Table 3).

After confirmation that the ERT protocol used successfully reducedprimary storage of α-Gal A substrates in Fabry knockout mice, weassessed the impact of ERT in correcting the secondary alteration thatwas found in our study. Indeed, ERT showed an effect on hepatic totalGlucCer levels. Total GlucCer was ≈1.9-fold elevated in the liver of1.5 mg/kg treated Fabry knockout mice relative to the group adminis-tered with saline solution (Fig. 3). In addition, the comparison of totalGlucCer levels between wild type (82.3 ± 20.3 μg/g tissue, Fig. 2A)and Fabry animals treated with 1.5 mg/kg α-Gal A (Fig. 3) indicatesthat the treatment protocol used completely normalized hepaticGlucCer levels. Treatment of Fabry animals with the lower dose of0.2 mg/kg α-Gal A did not result in a statistically significant increaseof hepatic GlucCer (Fig. 3). In contrast to what was observed in theliver, ERT had no effect on total GlucCer concentration in the other or-gans studied (Fig. 3).

These findings show a novel effect of ERT in Fabry knockoutmice: anorgan-specific increase of hepatic GlucCer levels.

3.4. Plasma from hemizygous Fabry patients has an altered GlucCer/Cer ra-tio that correlates with lyso-Gb3 concentration

We examined if the secondary alteration found in Fabry knockoutmice was also present in Fabry patients. The description of Fabry

Table 3Quantification of Gb3 and Lyso-Gb3 by UPLC–MS/MS in different tissues of treated and untrea

Untreated Treated

0.2 mg/kg α-

Tissue Gb3(μg/g tissue)

Lyso-Gb3(nmol/g tissue)

Gb3(μg/g tissue)

Spleen 3266 ± 810 13.6 ± 0.3 389 ± 36Liver 784 ± 239 8.5 ± 3.9 24 ± 16Kidney 1861 ± 425 1.7 ± 0.7 580 ± 47Heart 401 ± 63 1.4 ± 0.3 62 ± 18

Data are expressed asmean ± SD. nd—non detectable.n = 5untreated Fabry, except for liver a

patients' genetic, biochemical and clinical characteristics included inthis study is presented in Table 4. 15 patients had the F113L mutation,which has previously been associated with a mild cardiac phenotype(Eng et al., 1997), 1 had an E358del mutation previously associatedwith the classic phenotype (Blanch et al., 1996; Monserrat et al., 2007)and another patient had a M290I mutation which has also been report-ed previously (Shabbeer et al., 2006) and exhibited a clinical phenotypeof late-onset Fabry disease.

Total GlucCer, total Cer and respective isoformswere studied in plas-ma samples from Fabry patients and compared to gender-matched un-affected individuals. We did not observe any differences betweenhemizygous or heterozygous Fabry patients and the respective controlgroups (Supplemental Figs. A and B). Subsequently, we assessed the rel-ative amount of both lipids in individuals. This is particularly interestingbecause GlucCer and Cer can have opposing biological effects (see intro-duction). The study of plasma total GlucCer/total Cer ratio of individualFabry hemizygotes revealed a ≈20% increase comparatively to that ofunaffected males. In heterozygotes, no differences were found (Fig. 4A).

Previous studies suggest that plasma lyso-Gb3 levels could be a newbiomarker of Fabry disease (Rombach et al., 2010; Togawa et al., 2010).This prompted us to investigate if the increase in total GlucCer/total Cerratio of Fabry hemizygotes correlated with lyso-Gb3 concentration inFabry patients' plasma. A weak negative correlation between age andtotal GlucCer/total Cer ratio in unaffected individuals was observed(r = −0.37, p = 0.03; Supplemental Fig. C). To eliminate this effect,the variation in plasma total GlucCer/total Cer ratio was determinedfor Fabry patients relative to gender and age-comparable healthy sub-jects and plotted against the concentration of Lyso-Gb3 (Fig. 4B). InFabry hemizygotes, total GlucCer/total Cer variation showed a verystrong correlation with lyso-Gb3 concentration (r = 0.83, p = 0.008;Fig. 4B). In the case of heterozygotes, this correlation was absent(r = 0.17, p = 0.7; Fig. 4B). Two female patients (patients 12 and 14,Table 4) had lyso-Gb3 values below the detection limit (b4 nM). Patient12 had a later-onset Fabry phenotype and patient 14was asymptomaticat the time of analysis. Plasma lyso-Gb3 is a very useful marker in the

ted 12-weeks-old Fabry knockout mice.

Gal A 1.5 mg/kg α-Gal A

Lyso-Gb3(nmol/g tissue)

Gb3(μg/g tissue)

Lyso-Gb3(nmol/g tissue)

12.9 ± 0.0 404 ± 107 nd2.4 ± 0.3 4 ± 1 ndnd 144 ± 40 nd0.6 ± 0.2 16 ± 4 nd

ndkidney (n = 7). n = 5Fabry + 0.2 mg/kgα-Gal A; n = 3Fabry + 1.5 mg/kgα-Gal A.

Table 4Fabry patients' biochemical and phenotypical characteristics.

Patient Gender Age at sampling Mutation α-Gal A activity leukocytes(nmol/h/mg protein)

Phenotype

1 Male 40 F113L 2 Symptomatic (C)2 Male 47 F113L 3 Symptomatic (C)3 Male 51 F113L 2 Symptomatic (C + R + WML)4 Male 54 F113L 1 Symptomatic (C + R + WML)5 Male 60 F113L 2 Symptomatic (C + R + WML)6 Male 18 F113L 1 Asymptomatic7 Male 24 F113L 2 Asymptomatic8 Male 39 F113L 2 Asymptomatic9 Male 43 F113L 1 Symptomatic (C)10 Female 65 F113L 3 Symptomatic (C + WML)11 Female 50 E358del 27 Symptomatic (C)12 Female 47 M290I 18 Symptomatic (C)13 Female 49 F113L 18 Asymptomatic14 Female 69 F113L 56 Asymptomatic15 Female 40 F113L 29 Asymptomatic16 Female 64 F113L 50 Asymptomatic17 Female 52 F113L 27 Symptomatic (R)

C, cardiac involvement; R, renal involvement; WML, white matter lesions. Mean activity of controls in leukocytes (n = 123): 57 ± 14 nmol/h/mg protein.

102 R. Quinta et al. / Gene 536 (2014) 97–104

discrimination of male Fabry patients from healthy controls, but it canbe undetectable in female heterozygotes, particularly those with alate-onset phenotype (Togawa et al., 2010; van Breemen et al., 2011).Although the overall correlation coefficient (including male and femalepatients, carrying the F113L, M290I and E358del mutations) was statis-tically significant (r = 0.56, p = 0.02), this result was clearly influ-enced by the male observations that counterbalance the very weakcorrelation observed in females. The same reasoning can be madewith respect to the overall group of 15 patients carrying the F113L mu-tation (9males and 8 females, r = 0.77, p b 0.001) since, as observed inFig. 4B, the correlation in F113L heterozygotes is very weak.

4. Discussion

The intralysosomal accumulation of unmetabolized substrates insphingolipidoses originates a wide array of clinical phenotypes. Thisclinical heterogeneity does not have a clear association with the levelof enzymatic activity, suggesting that many secondary biochemicalor cellular pathways must also be affected (Futerman and vanMeer, 2004; Prinetti et al., 2011). Secondary sphingolipid alterationsin Fabry disease and other sphingolipidoses have been reported(Chen et al., 1999; Fuller et al., 2005; Meikle et al., 2008). In thepresent study, we evaluated if α-Gal A deficiency has an effect onGlucCers and Cers.

Our results showed that α-Gal A deficiency in mice causes a reduc-tion in GlucCer level. A marked decrease in GlucCer in mouse aortic en-dothelial cells (MAECs) isolated from young Fabry knockout mice was

Fig. 4. Quantification of GlucCer/Cer ratio and association with lyso-Gb3 concentration in plassamples from unaffected individuals and Fabry patients. Data are expressed asmean ± SD. (B)of Fabry patients. This variation was determined as a percentage relative to the mean values ofsents a different individual. *p b 0.05. Males: controls n = 20, patients n = 9. Females: contro

also observed in a previous study by Shu and coworkers (Shu andShayman, 2007). Our study reveals that this reduction is not restrictedto MAECs. Accordingly, Fuller et al. (2005) observed a similar decreaseof GlucCer in the urine of Fabry heterozygotes. The decrease in GlucCeris most likely explained by the fact that a significant amount of GSLs isnot recycling, as it is trapped as Gb3. This idea is confirmed by the factthat ERT, which increases degradation of Gb3, allowing it to be recycled,also results in normalization of GlucCer levels. Another possibility is thatthe blockage of proper catabolism at the lysosomes in sphingolipidoses,such as Fabry disease, can lead to a jamof the overall flowofmetabolites(Prinetti et al., 2011; Simons and Gruenberg, 2000). As such, the prima-ry accumulation of α-Gal A substrates could disturb intracellular trans-port of GlucCer. The decreased level of GlucCer could also be the resultof an active homeostatic response (Fuller, 2010), in an attempt to re-store the sphingolipid balance that was compromised by impaired deg-radation of Gb3.

It is not clear at themoment if the decrease in GlucCer by itself has anegative biological impact. GlucCer is known to have pro-mitogenic ef-fects in a variety of cell types (Datta and Radin, 1988;Marsh et al., 1995;Shayman et al., 1991) and increases in its level are known to be associ-atedwithmany types of cancer cellswith increased drug resistance. De-creasingGlucCer throughGlucosylceramide synthase inhibition in thesecells has been shown to increase the intracellular level of Cer, decreasingdrug resistance (Kartal Yandim et al., 2013). Importantly, the decreasein GlucCer in Fabry tissues was not accompanied by statistically signifi-cant increases in Cer, which is known to be a bioactive sphingolipid pro-moting apoptosis and tissue damage.

ma of Fabry patients as analyzed by UPLC–MS/MS. (A) GlucCer/Cer ratio quantification inGlucCer/Cer ratio variation was plotted against lyso-Gb3 concentration in plasma samples2–3 age-comparable (±3 years) and gender-matched healthy subjects. Each point repre-ls n = 12, patients n = 8.

103R. Quinta et al. / Gene 536 (2014) 97–104

We also report that GlucCer isoforms are unequally reduced in dif-ferent tissues of Fabry knockout mice. The biological roles of differentsphingolipid isoforms are starting to be elucidated. GlucCer was initiallyreported to be non-antigenic for iNKT cells (Stanic et al., 2003). It is nowknown that GlucCer is, in fact, an iNKT cell antigen but its activity de-pends on the composition of the acyl chain. GlucCer isoforms 24:1 and18:1, but not 16:0 or 18:0, are human and mouse iNKT cells antigensand the 24:1 isoform is a particularly potent antigen (Brennan et al.,2011). iNKT cells are known to be numerically decreased in Fabryknockout mice but not in Fabry patients (Balreira et al., 2008). In addi-tion, our group has also demonstrated that iNKT cells are unequally af-fected in different organs of these mice (Macedo et al., 2012). Ourfindings that GlucCer 24:1 is reduced in major peripheral organs ofiNKT homing, the liver and spleen,may inspire additional studies to dis-cover the underlying mechanisms of these immune system alterationsin the Fabry disease mouse model.

The organ-specific effect of ERT on the normalization of hepaticGlucCer level is probably the result of a more efficient correction ofα-Gal A deficiency in the liver, as compared to other organs ofFabry knockout mice. As expected, ERT drastically reduced Gb3 andlyso-Gb3 storage in hepatic cells, consistent with previous reportsof a higher α-Gal A uptake in the liver (Ioannou et al., 2001; Leeet al., 2003). Of interest, we noted that independent of theα-Gal A dos-age used (0.2 mg/kg or 1.5 mg/kg), Gb3 levels in the spleen of treatedmice were also highly reduced by ERT (≈90%). Moreover, by measure-ment of α-Gal A activity, we noted that a high amount of α-Gal A waspresent in the spleen of treated Fabry knockout mice, although thislevel was significantly lower than in the liver (data not shown). Thesefindings suggest a relatively high splenic uptake of α-Gal A, probablydue to the high abundance of macrophages in this organ, and residuallevels of Gb3 are possibly caused by the presence of red blood cells.However, ERT was unable to correct the secondary decrease in GlucCerin the spleen. A possible explanation for this observation might be thatnot all splenic cells are capable of endocytosing sufficient recombinantenzyme to correct this secondary alteration, and it would be interestingto further explore this hypothesis.

Fuller et al. (2005) analyzed the urine of Fabry patients and reportedsecondary increases in Cer 16:0 and 24:1 in hemizygotes and decreasesin GlucCer 22:0 and 24:1 in heterozygotes. In the present study, we didnot observe any alterations in these or other Cer or GluCer isoforms inthe plasma of Fabry patients. This observation raises two importantquestions. First, these alterations might be tissue/cell-dependent. Infact, Fuller and coworkers suggest that the differences observed intheir study might be the consequence of a specific disrupted lipid flowin renal epithelial cells. Secondly, the differences between both studiescan be related with distinct genetic and phenotypical characteristics ofthe Fabry populations analyzed.

Intriguingly, and in contrast to what was observed in the Fabry dis-ease mouse model, individual GlucCer/Cer ratio in plasma samples ofFabry hemizygotes is slightly (≈20%) elevated comparatively to thatof unaffected individuals. GSLs, including GlucCer and Cer, are compo-nents of VLDL, LDL and HDL lipoprotein particles (Chatterjee et al.,1984; Hammad et al., 2010). GSLs in VLDL and LDL particles largelystem from the novel synthesis in the liver. In contrast, HDL particleslargely derive from tissues other than the liver. Men andmice differ sig-nificantly in plasma profile of lipoprotein particles. VLDL and LDL arepresent in large quantities in human plasma while HDL is the predomi-nant lipoprotein particle in mice (Chapman, 1986). Consequently, plas-ma lipids in mice reflect better the situation in tissue as compared tohumanplasma. This could account for the observed differences betweenplasmas of both species.

Our results also showed that plasmaGlucCer/Cer ratiowas normal inFabry female patients. Primary lipid alterations in plasma Gb3 and lyso-Gb3 are known to be more prominent in Fabry hemizygotes with theclassic phenotype than in heterozygotes (Togawa et al., 2010; Vedderet al., 2007). It is reasonable that this imbalance follows the pattern

observed for primary lipid storage, being increased in hemizygotes butabsent in heterozygotes. The abnormalities in plasma GlucCer/Cerratio were found to correlate with lyso-Gb3 concentration in plasmasamples of Fabry hemizygotes, suggesting that this ratio may hold par-ticular promise for improved assessment of these patients. However, tofully validate this hypothesis, more samples from Fabry patients, wellcharacterized at genetic and clinical levels, and covering more differentgenotypes/phenotypes need to be analyzed.

It remains unclearwhether secondary sphingolipid changes in Fabrydisease are pathological events that contribute to compromise cell func-tions or homeostatic responses. Fully understanding the underlyingmechanisms of secondary alterations and their biological significancewill be an important factor to clarify the pathophysiology of Fabrydisease. Future studies addressing the links between clinical manifesta-tions, GlucCer, Cer, lyso-Gb3 and other bioactive sphingolipid metabo-lites might contribute to the identification of new biomarkers.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.gene.2013.11.073.

Conflict of interest statement

The authors declare that they have no conflicts of interest.

Role of funding source

This work is funded by FEDER funds through the Operational Com-petitiveness Programme—COMPETE and by National Funds throughFCT—Fundação para a Ciência e a Tecnologia under the project FCOMP-01-0124-FEDER-022718 (PEst-C/SAU/LA0002/2011). The sponsors hadno influence on the conception, study design, data collection, analysisor interpretation,manuscriptwriting aswell as on the decision to submitthe article for publication. R Quinta was supported by a fellowship fromFundação para a Ciência e Tecnologia (SFRH/BD/33447/2008). This workis part of the PhD thesis of R Quinta.

Acknowledgments

The authors would like to thank Professor R.O. Brady for kindly pro-viding the Fabry knockout mice, Shire Human Genetics for agalsidasealfa (Replagal©), Dr. Lorena G Rodrigues for critical discussions andmanagement of mouse colonies and Prof. Dr. Johannes Aerts for criticaldiscussions.

References

Aerts, J.M., et al., 2008. Elevated globotriaosylsphingosine is a hallmark of Fabry disease.Proc. Natl. Acad. Sci. U. S. A. 105, 2812–2817.

Auray-Blais, C., et al., 2010. How well does urinary lyso-Gb3 function as a biomarker inFabry disease? Clin. Chim. Acta 411, 1906–1914.

Balreira, A., et al., 2008. Anomalies in conventional T and invariant natural killer T-cellpopulations in Fabry mice but not in Fabry patients. Br. J. Haematol. 143, 601–604.

Banikazemi, M., et al., 2007. Agalsidase-beta therapy for advanced Fabry disease: a ran-domized trial. Ann. Intern. Med. 146, 77–86.

Bekri, S., Lidove, O., Jaussaud, R., Knebelmann, B., Barbey, F., 2006. The role of ceramidetrihexoside (globotriaosylceramide) in the diagnosis and follow-up of the efficacyof treatment of Fabry disease: a review of the literature. Cardiovasc. Hematol. AgentsMed. Chem. 4, 289–297.

Bielawski, J., Szulc, Z.M., Hannun, Y.A., Bielawska, A., 2006. Simultaneous quantitativeanalysis of bioactive sphingolipids by high-performance liquid chromatography–tandem mass spectrometry. Methods 39, 82–91.

Blanch, L.C., Meaney, C., Morris, C.P., 1996. A sensitive mutation screening strategy forFabry disease: detection of nine mutations in the alpha-galactosidase A gene. Hum.Mutat. 8, 38–43.

Brennan, P.J., et al., 2011. Invariant natural killer T cells recognize lipid self antigen in-duced by microbial danger signals. Nat. Immunol. 12, 1202–1211.

Chapman, M.J., 1986. Comparative analysis of mammalian plasma lipoproteins. MethodsEnzymol. 128, 70–143.

Chatterjee, S., Bell, W.R., Kwiterovich Jr., P.O., 1984. Distribution of antithrombin III andglucosylceramide in human plasma lipoproteins and lipoprotein deficient plasma.Lipids 19, 363–366.

Chen, C.S., Patterson, M.C., Wheatley, C.L., O'Brien, J.F., Pagano, R.E., 1999. Broad screeningtest for sphingolipid-storage diseases. Lancet 354, 901–905.

104 R. Quinta et al. / Gene 536 (2014) 97–104

Ciron, C., et al., 2006. Gene therapy of the brain in the dog model of Hurler's syndrome.Ann. Neurol. 60, 204–213.

Cressant, A., et al., 2004. Improved behavior and neuropathology in the mouse model ofSanfilippo type IIIB disease after adeno-associated virus-mediated gene transfer inthe striatum. J. Neurosci. 24, 10229–10239.

Datta, S.C., Radin, N.S., 1988. Stimulation of liver growth and DNA synthesis byglucosylceramide. Lipids 23, 508–510.

Desmaris, N., Verot, L., Puech, J.P., Caillaud, C., Vanier, M.T., Heard, J.M., 2004. Preven-tion of neuropathology in the mouse model of Hurler syndrome. Ann. Neurol. 56,68–76.

Durant, B., et al., 2011. Sex differences of urinary and kidney globotriaosylceramide andlyso-globotriaosylceramide in Fabry mice. J. Lipid Res. 52, 1742–1746.

Ellinwood, N.M., et al., 2007. Bone marrow transplantation for feline mucopolysaccharidosisI. Mol. Genet. Metab. 91, 239–250.

Eng, C.M., Ashley, G.A., Burgert, T.S., Enriquez, A.L., D'Souza, M., Desnick, R.J., 1997. Fabrydisease: thirty-fivemutations in the alpha-galactosidase A gene in patients with clas-sic and variant phenotypes. Mol. Med. 3, 174–182.

Frisella, W.A., et al., 2001. Intracranial injection of recombinant adeno-associated virusimproves cognitive function in a murine model of mucopolysaccharidosis type VII.Mol. Ther. 3, 351–358.

Fuller, M., 2010. Sphingolipids: the nexus between Gaucher disease and insulin resis-tance. Lipids Health Dis. 9, 113.

Fuller, M., et al., 2005. Urinary lipid profiling for the identification of Fabry hemizygotesand heterozygotes. Clin. Chem. 51, 688–694.

Futerman, A.H., van Meer, G., 2004. The cell biology of lysosomal storage disorders. Nat.Rev. Mol. Cell Biol. 5, 554–565.

Germain, D.P., 2010. Fabry disease. Orphanet J. Rare Dis. 5, 30.Griffin, L.D., Gong, W., Verot, L., Mellon, S.H., 2004. Niemann–Pick type C disease involves

disrupted neurosteroidogenesis and responds to allopregnanolone. Nat. Med. 10,704–711.

Hammad, S.M., et al., 2010. Blood sphingolipidomics in healthy humans: impact of samplecollection methodology. J. Lipid Res. 51, 3074–3087.

Hannun, Y.A., Obeid, L.M., 2008. Principles of bioactive lipid signalling: lessons fromsphingolipids. Nat. Rev. Mol. Cell Biol. 9, 139–150.

Ioannou, Y.A., Zeidner, K.M., Gordon, R.E., Desnick, R.J., 2001. Fabry disease: preclinicalstudies demonstrate the effectiveness of alpha-galactosidase A replacement inenzyme-deficient mice. Am. J. Hum. Genet. 68, 14–25.

Kartal Yandim, M., Apohan, E., Baran, Y., 2013. Therapeutic potential of targetingceramide/glucosylceramide pathway in cancer. Cancer Chemother. Pharmacol. 71,13–20.

Lee, K., et al., 2003. A biochemical and pharmacological comparison of enzyme replace-ment therapies for the glycolipid storage disorder Fabry disease. Glycobiology 13,305–313.

Macedo, M.F., Quinta, R., Pereira, C.S., Sa Miranda, M.C., 2012. Enzyme replacement ther-apy partially prevents invariant Natural Killer T cell deficiency in the Fabry diseasemouse model. Mol. Genet. Metab. 106, 83–91.

Marsh, N.L., Elias, P.M., Holleran,W.M., 1995. Glucosylceramides stimulate murine epider-mal hyperproliferation. J. Clin. Invest. 95, 2903–2909.

Meikle, P.J., Hopwood, J.J., Clague, A.E., Carey, W.F., 1999. Prevalence of lysosomal storagedisorders. JAMA 281, 249–254.

Meikle, P.J., et al., 2008. Effect of lysosomal storage on bis(monoacylglycero)phosphate.Biochem. J. 411, 71–78.

Monserrat, L., et al., 2007. Prevalence of Fabry disease in a cohort of 508 unrelated pa-tients with hypertrophic cardiomyopathy. J. Am. Coll. Cardiol. 50, 2399–2403.

Mullen, T.D., et al., 2011. Selective knockdown of ceramide synthases reveals complexinterregulation of sphingolipid metabolism. J. Lipid Res. 52, 68–77.

Ohshima, T., et al., 1999. Aging accentuates and bonemarrow transplantation amelioratesmetabolic defects in Fabry disease mice. Proc. Natl. Acad. Sci. U. S. A. 96, 6423–6427.

Poorthuis, B.J., et al., 1999. The frequency of lysosomal storage diseases in TheNetherlands. Hum. Genet. 105, 151–156.

Prinetti, A., Prioni, S., Chiricozzi, E., Schuchman, E.H., Chigorno, V., Sonnino, S., 2011. Sec-ondary alterations of sphingolipid metabolism in lysosomal storage diseases.Neurochem. Res. 36, 1654–1668.

Rodrigues, L.G., et al., 2009. Neurophysiological, behavioral and morphological abnormal-ities in the Fabry knockout mice. Neurobiol. Dis. 33, 48–56.

Rombach, S.M., et al., 2010. Plasma globotriaosylsphingosine: diagnostic value and rela-tion to clinical manifestations of Fabry disease. Biochim. Biophys. Acta 1802,741–748.

Schiffmann, R., Ries, M., Timmons, M., Flaherty, J.T., Brady, R.O., 2006. Long-term therapywith agalsidase alfa for Fabry disease: safety and effects on renal function in a homeinfusion setting. Nephrol. Dial. Transplant. 21, 345–354.

Schiffmann, R., et al., 2013. Changes in plasma and urine globotriaosylceramide levels do notpredict Fabry disease progression over 1 year of agalsidase alfa. Genet. Med. 15, 983–989.

Schwab, U., et al., 2008. Triacylglycerol fatty acid composition in diet-induced weight lossin subjects with abnormal glucose metabolism—the GENOBIN study. PLoS One 3,e2630.

Shabbeer, J., Yasuda, M., Benson, S.D., Desnick, R.J., 2006. Fabry disease: identification of 50novel alpha-galactosidase A mutations causing the classic phenotype and three-dimensional structural analysis of 29 missense mutations. Hum. Genomics 2, 297–309.

Shayman, J.A., et al., 1991. Modulation of renal epithelial cell growth by glucosylceramide.Association with protein kinase C, sphingosine, and diacylglycerol. J. Biol. Chem. 266,22968–22974.

Shu, L., Shayman, J.A., 2007. Caveolin-associated accumulation of globotriaosylceramidein the vascular endothelium of alpha-galactosidase A null mice. J. Biol. Chem. 282,20960–20967.

Simons, K., Gruenberg, J., 2000. Jamming the endosomal system: lipid rafts and lysosomalstorage diseases. Trends Cell Biol. 10, 459–462.

Spada, M., et al., 2006. High incidence of later-onset Fabry disease revealed by newbornscreening. Am. J. Hum. Genet. 79, 31–40.

Stanic, A.K., et al., 2003. Defective presentation of the CD1d1-restricted natural Va14Ja18NKT lymphocyte antigen caused by beta-D-glucosylceramide synthase deficiency.Proc. Natl. Acad. Sci. U. S. A. 100, 1849–1854.

Togawa, T., et al., 2010. Plasma globotriaosylsphingosine as a biomarker of Fabry disease.Mol. Genet. Metab. 100, 257–261.

van Breemen, M.J., et al., 2011. Reduction of elevated plasma globotriaosylsphingosine inpatients with classic Fabry disease following enzyme replacement therapy. Biochim.Biophys. Acta 1812, 70–76.

Vedder, A.C., et al., 2007. The Dutch Fabry Cohort: diversity of clinical manifestations andGb3 levels. J. Inherit. Metab. Dis. 30, 68–78.

Walkley, S.U., Vanier, M.T., 2009. Secondary lipid accumulation in lysosomal disease.Biochim. Biophys. Acta 1793, 726–736.

Wang, R.Y., Lelis, A., Mirocha, J., Wilcox, W.R., 2007. Heterozygous Fabry women are notjust carriers, but have a significant burden of disease and impaired quality of life.Genet. Med. 9, 34–45.

Whitfield, P.D., et al., 2005. Monitoring enzyme replacement therapy in Fabry disease—role of urine globotriaosylceramide. J. Inherit. Metab. Dis. 28, 21–33.

Wittmann, J., et al., 2012. Newborn screening for lysosomal storage disorders in Hungary.JIMD Rep. 6, 117–125.

Wu, D., et al., 2007. Aging up-regulates expression of inflammatory mediators in mouseadipose tissue. J. Immunol. 179, 4829–4839.

Young, E., et al., 2005. Is globotriaosylceramide a useful biomarker in Fabry disease? ActaPaediatr. Suppl. 94, 51–54.