Inhibition of angiogenesis by  -galactosylceramidase deficiency in globoid cell leukodystrophy

doi:10.1093/brain/awt215

Brain 2013: 136; 2859–2875

| 2859

BRAIN A JOURNAL OF NEUROLOGY

Inhibition of angiogenesis by b-galactosylceramidase deficiency in globoid cell leukodystrophy

1 2 3 4

Department of Molecular and Translational Medicine, School of Medicine, University of Brescia, Italy Department of Basic Biomedical Sciences, Unit of Human Anatomy and Histology, University of Bari, Italy National Cancer Institute, Giovanni Paolo II, Bari, Italy Department of Anatomy and Cell Biology, University of Illinois, Chicago, IL, USA

Correspondence to: Marco Presta, Department of Molecular and Translational Medicine, Viale Europa 11, 25123 Brescia, Italy E-mail: [email protected]

Globoid cell leukodystrophy (Krabbe disease) is a neurological disorder of infants caused by genetic deficiency of the lysosomal enzyme b-galactosylceramidase leading to accumulation of the neurotoxic metabolite 1-b-D-galactosylsphingosine (psychosine) in the central nervous system. Angiogenesis plays a pivotal role in the physiology and pathology of the brain. Here, we demonstrate that psychosine has anti-angiogenic properties by causing the disassembling of endothelial cell actin structures at micromolar concentrations as found in the brain of patients with globoid cell leukodystrophy. Accordingly, significant alterations of microvascular endothelium were observed in the post-natal brain of twitcher mice, an authentic model of globoid cell leukodystrophy. Also, twitcher endothelium showed a progressively reduced capacity to respond to pro-angiogenic factors, defect that was corrected after transduction with a lentiviral vector harbouring the murine b-galactosylceramidase complementary DNA. Finally, RNA interference-mediated b-galactosylceramidase gene silencing causes psychosine accumulation in human endothelial cells and hampers their mitogenic and motogenic response to vascular endothelial growth factor. Accordingly, significant alterations were observed in human microvasculature from brain biopsy of a globoid cell leukodystrophy case. Together these data demonstrate that b-galactosylceramidase deficiency induces significant alterations in endothelial neovascular responses that may contribute to central nervous system and systemic damages that occur in globoid cell leukodystrophy.

Keywords: angiogenesis; Krabbe disease; neurodegeneration; psychosine; twitcher mice Abbreviations: FGF2-T-MAE = FGF2-overexpressing murine aortic endothelial cells; glucopsychosine = 1-b-D-glucosylsphingosine; psychosine = 1-b-D-galactosylsphingosine

Introduction Lysosomal storage disorders represent one of the most frequent classes of human genetic diseases. They are characterized by the accumulation of disease-specific metabolic intermediates within

lysosomes, leading to severe organ damage and premature death (Ballabio and Gieselmann, 2009). Globoid cell leukodystrophy, or Krabbe disease, is an autosomal recessive sphingolipidosis caused by genetic deficiency of the lysosomal hydrolase b-galactosylceramidase (GALC)

Received January 23, 2013. Revised May 17, 2013. Accepted June 12, 2013 ß The Author (2013). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://brain.oxfordjournals.org/ at Università degli Studi di Brescia on September 26, 2013

Mirella Belleri,1 Roberto Ronca,1 Daniela Coltrini,1 Beatrice Nico,2 Domenico Ribatti,2,3 Pietro L. Poliani,1 Arianna Giacomini,1 Patrizia Alessi,1 Sergio Marchesini,1 Marta B. Santos,4 Ernesto R. Bongarzone4 and Marco Presta1

2860

| Brain 2013: 136; 2859–2875

(Suzuki and Suzuki, 1970). The disease is characterized by degeneration of oligodendroglia and progressive demyelination of the peripheral and CNS. Clinically, globoid cell leukodystrophy manifests in early infancy and results in a severe neurological dysfunction that often leads to death by 2 years of age (Loonen et al., 1985; Wenger et al., 2000; Suzuki, 2003). GALC degrades galactosylceramide (a major component of myelin) and other terminal b-galactose-containing sphingolipids, including b-galactosylsphingosine (psychosine). At present, the only clinical treatment for globoid cell leukodystrophy is bone marrow or umbilical cord blood cell transplantation for late-onset and presymptomatic patients (Wenger et al., 2000; Sakai, 2009). Thus, understanding the molecular pathogenesis of globoid cell leukodystrophy remains a high priority from a clinical standpoint. The pathogenesis of globoid cell leukodystrophy has been proposed to arise from the accumulation of psychosine, the neurotoxic lysolipid metabolite detected at high levels in the CNS of patients with globoid cell leukodystrophy when compared with healthy individuals (Igisu and Suzuki, 1984; Suzuki, 1998; Ballabio and Gieselmann, 2009). Accumulation of psychosine leads to cytotoxic effects on oligodendroglial cells, triggering apoptotic cell death in vitro and in vivo (Jatana et al., 2002; Haq et al., 2003; Zaka and Wenger, 2004) and alterations in membrane architecture (White et al., 2009, 2011). Also, psychosine affects various enzymes involved in signal transduction pathways (Ballabio and Gieselmann, 2009) and axonal cytoskeleton and transport (Castelvetri et al., 2011; Smith et al., 2011; Cantuti-Castelvetri et al., 2012), and hampers actin reorganization leading to the formation of microglia and macrophage-derived multinuclear globoid cells (Kanazawa et al., 2000; Kozutsumi et al., 2002; Ijichi et al., 2013). Neovascularization plays an important role in the development of the CNS and protects it from neurological disorders (Greenberg and Jin, 2005; Segura et al., 2009). The tight cross-talk among glial, neuronal and endothelial cells in CNS is underlined by the capacity of angiogenic factors, including vascular endothelial growth factor-A (VEGF) and fibroblast growth factor-2 (FGF2), to modulate neurogenesis and neuroprotection. In turn, neurotrophic factors may regulate angiogenesis. These common classes of molecules affecting both vascular and neuronal functions have been recently termed ‘angioneurins’ (Zacchigna et al., 2008). As a consequence of this neurovascular cross-talk, blood vessel alterations appear to be implicated in the pathogenesis of stroke and neurodegenerative disorders, including Alzheimer’s disease, Parkinson’s disease, multiple sclerosis and amyotrophic lateral sclerosis (Greenberg and Jin, 2005; Zlokovic, 2005; Zacchigna et al., 2008; Segura et al., 2009). Moreover, sporadic pathological studies of the brain vasculature on a limited number of patients indicate that alterations of the angioarchitecture of brain microvasculature may occur also in human leukodystrophies (Kondo and Suzuki, 1993; Buee et al., 1994). These findings prompted us to investigate the effect of GALC deficiency and psychosine accumulation on CNS microvascularization and the angiogenic process. The data demonstrate that psychosine is endowed with anti-angiogenic activity by causing actin cytoskeleton disassembly in endothelial cells. Accordingly, significant microvascular alterations occur in the brain of twitcher mice,

M. Belleri et al. an authentic model of globoid cell leukodystrophy that develops most of the pathological hallmarks seen in human patients, including psychosine accumulation (Kobayashi et al., 1980; Suzuki, 1995). Also, the endothelium of juvenile twitcher mice shows a progressively reduced capacity to respond in vivo or ex vivo to exogenously administered pro-angiogenic factors. Normalization of twitcher endothelium was observed after in vitro murine Galc complementary DNA transduction. Finally, downregulation of GALC activity following lentivirus-mediated short hairpin RNA silencing causes psychosine accumulation and reduces the proliferative capacity and motogenic activity of human umbilical vein endothelial cells. In keeping with these observations, significant alterations were observed in the brain microvasculature of a human globoid cell leukodystrophy biopsy. In conclusion, our data indicate that GALC deficiency may induce significant alterations of the angiogenic process and CNS vascularization. In turn, these alterations may contribute to CNS and systemic damages that occur in globoid cell leukodystrophy.

Materials and methods Details beyond the descriptions provided here are given in the online Supplementary material.

Reagents Psychosine from bovine brain with a length of sphingoid base of C18 carbon atoms (molecular weight: 461.63, purity 598%), glucopsychosine (1-b-D-glucosylsphingosine) from glucocerebrosides from human Gaucher’s spleen and N-acetyl-D-sphingosine (molecular weight: 341.53, purity 598%) were purchased from Sigma Chemical Co. Stock solutions were prepared in dimethyl sulphoxide and diluted in cell culture medium to a final dimethyl sulphoxide concentration 4 0.5% (vol:vol); the same doses of vehicle were ineffective in all of the assays.

Cell cultures Murine aortic endothelial cells and brain microvascular endothelial cells were grown in Dulbecco’s modified Eagle’s medium with 10% foetal calf serum. FGF2-overexpressing murine aortic endothelial (FGF2T-MAE) cells (Gualandris et al., 1996) were grown in Dulbecco’s modified Eagle’s medium supplemented with 4 mM glutamine and 10% foetal calf serum. Murine microvascular lung endothelial 1G11 cells and subcutaneous microvascular endothelial cells (Dong et al., 1997) were grown in Dulbecco’s modified Eagle’s medium supplemented with 1 mM glutamine, 1% non-essential amino acids, 1 mM sodium pyruvate and 20% foetal calf serum in the presence of 10 ng/ml FGF2 and 10 mg/ml heparin. Bovine aortic endothelial cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% heatinactivated donor calf serum. Human umbilical vein endothelial cells were grown in M199 medium supplemented with 20% foetal calf serum, 100 mg/ml endothelial cell growth factor and 100 mg/ml heparin.

Cell proliferation assays Murine aortic endothelial, FGF2-T-MAE and bovine aortic endothelial cells were seeded at 20 000 cells/cm2 whereas 1G11 cells were seeded

GALC deficiency and angiogenesis

Brain 2013: 136; 2859–2875

| 2861

at 50 000 cells/cm2. After overnight incubation, cells were incubated with increasing concentrations of psychosine for 24 h. Cells were then trypsinized and counted in a Burker chamber.

Brain expression of angiogenic growth factors

Wounding of endothelial cell monolayers

Murine brains were analysed for the expression of different proangiogenic factors by quantitative reverse transcriptase PCR and data were normalized for Gapdh and Actin expression (Coltrini et al., 2013) using primers listed in Supplementary Table 1.

Wounds were created in 1G11 or murine aortic endothelial cell monolayers with a 1.0-mm wide rubber policeman. Cells were incubated in fresh medium with 10% foetal calf serum and increasing concentrations of psychosine. At different time points, the percentage of cells at the edge of the wound showing cell membrane ruffles and repositioning of the microtubule organization centre were evaluated (Belleri et al., 2005).

b-Galactosylceramidase enzyme assay GALC-mediated lysis of the synthetic fluorescent GALC substrate LRh-6-GalCer (N-lissamine rhodaminyl-6-aminohexanoylgalactosyl ceramide) following its incubation with cell pellets (50 mg) was quantified by thin-layer chromatography as described previously (Marchesini et al., 1990).

Morphometric evaluation of murine brain vascular density Sections (8 mm) of paraformaldehyde-fixed brain cortex and periventricular area samples were immunostained with anti-factor VIII-related antigen/von Willebrand’s factor antibody (FVIII/vWF) and counterstained with Gill’s haematoxylin. FVIII/vWF + vascular density was assessed on 30 sections per brain. Four to six  200 fields covering almost the whole of each of four sections per sample were examined with a square reticulum (0.78 mm2) inserted in the eyepiece. Means  1 standard deviation (SD) were determined for each section, sample and group of samples.

Three dimensional gel invasion assays

Three dimensional rendering of brain microvasculature

FGF2-T-MAE cell aggregates embedded in fibrin gel (Gualandris et al., 1996) were treated with psychosine in the presence of 10 mg/ml aprotinin to prevent the dissolution of the substrate. Formation of cell sprouts was quantified after 24 h by computerized image analysis. Confluent bovine aortic endothelial cells seeded on type I-collagen gel were treated with FGF2 plus VEGF (both at 30 ng/ml) with 10% foetal calf serum, either with or without psychosine. After 24 h, cells invading the gel in a plane of focus beneath the cell monolayer surface were quantified by computerized image analysis.

Free-floating 50-mm sections of paraformaldehyde-fixed, agaroseembedded brains from Day 36 wild-type and homozygous mice were incubated overnight at 4 C with an anti-CD31 antibody. After further incubation with secondary Alexa FluorÕ 594-conjugated antibody, sections were examined under a Zeiss Axiovert 200 M microscope equipped with ApoTome optical sectioning device. Z-stacks images were acquired with a  20 objective at 0.32 pixels/mm resolution, processed with AxioVision 4.8 software and 3D renderings were obtained using ImageJ software.

Endothelial cell adhesion and spreading

Confocal immunofluorescence microscopy

Bovine aortic endothelial cells were seeded at 25 000/cm2 on fibronectin-coated coverslips in the presence of vehicle or 50 mM psychosine. At different time points, adherent cells were stained with rhodaminephalloidin and the percentage of adherent cells, cells with actin disassembly and cells showing a spread morphology were counted.

Chick embryo chorioallantoic membrane assay

Cryosections (12 mm) from wild-type and twitcher brains were exposed overnight at 4 C to mouse anti-GFAP monoclonal antibody and rabbit anti-ZO-1 polyclonal antibody. Sections were incubated with Alexa Fluor 488 anti-mouse and Alexa FluorÕ 555 anti-rabbit antibodies. After nuclear staining, sections were examined under a Leica TCS SP2 confocal laser scanning microscope.

Transmission electron microscopy

Gelatine sponges (n = 8) containing vehicle or 500 ng of FGF2 with or without psychosine (1.6 mmoles) were placed on chicken embryo chorioallantoic membrane at Day 8 (Ribatti et al., 2006). At Day 12, blood vessels converging toward the implant were counted by two observers in a double-blind fashion.

Glutaraldehyde-fixed brain specimens were post-fixed with 1% osmium tetroxide and embedded in Epon 812. Ultrathin sections (60 nm) were stained with uranyl acetate/lead citrate and examined with a Zeiss EM109 electron microscope.

Mice

Evaluation of brain vascular permeability

Breeder twitcher heterozygous mice (C57BL/6J; Jackson Laboratories) were maintained under standard housing conditions. Animal handling protocols were in accordance with Italian institutional guidelines for animal care and use. Twitcher mutation was determined by PCR on DNA extracted from clipped tails (Sakai et al., 1996). Littermate wildtype (wt), heterozygous carrier (twi/ + ) and homozygous (homozygous) animals were used in all of the experiments.

Five minutes after intracardiac injection of sulphosuccinimidyl6-(biotinamido) hexanoate (sulpho-NHS-LC-biotin) (Rybak et al., 2005), brains were fixed in 4% paraformaldehyde and embedded in 6% agarose. Sections (100 mm) were stained with Alexa FluorÕ 488conjugated streptavidin. Z-stacks images were acquired with a  10 objective at 0.65 pixels/mm resolution using a Axiovert 200 M

2862

| Brain 2013: 136; 2859–2875

microscope equipped with ApoTome optical sectioning device and processed with AxioVision 4.8 software.

b-galactosylceramidase immunofluorescence analysis Cryosections (7 mm) of brain cortex were incubated overnight at 4 C with an anti-GALC polyclonal antibody followed by 1 h incubation with Alexa Fluor 488 anti-rabbit antibody. Sections were then incubated for 2 h with biotin-conjugated Bandeiraea simplicifolia BSI-B4 lectin, followed by 1 h incubation with streptavidin Alexa FluorÕ 594. Images were taken using a Zeiss LSM 510 META confocal laser scanning microscope.

Matrigel plug angiogenesis assay C57BL mice were injected subcutaneously with 400 ml MatrigelTM (Trevigen) containing PBS or 300 ng FGF2 in the absence or in the presence of 200 mM psychosine. In a second set of experiments, wildtype, heterozygous carrier (twi/ + ) and homozygous (homozygous) mice were injected subcutaneously with 400 ml MatrigelTM containing PBS or 500 ng FGF2 plus 100 ng VEGF. After 7 days, the vascular response was quantified by anti-CD31 immunostaining of frozen sections of the plugs (five sections/plug) followed by computerized image analysis of CD31 + areas in six microscopic fields per section. Also, the levels of CD31 expression in MatrigelTM plugs was evaluated by quantitative reverse transcriptase PCR and normalized for Gapdh expression (Coltrini et al., 2013). The primers are listed in Supplementary Table 1.

Murine aorta ring assay This was performed on fibrin-embedded aortic rings (Ronca et al., 2010) incubated with serum-free endothelial cell basal medium plus 10 mg/ml aprotinin in the presence of 30 ng/ml VEGF. When indicated, aorta rings were incubated in serum-free medium with 8 mg/ml polybrene and lentiviral particles harbouring the murine GALC complementary DNA or control EGFP complementary DNA (provided by A. Biffi, San Raffele Scientific Institute, Milan). After 24 h, rings were embedded in fibrin gel and incubated with VEGF in the presence of 10% foetal calf serum. Vessel sprouts were counted under a stereomicroscope at 5 days. Histological sections of aorta rings were stained with haematoxylin and eosin or decorated with anti-CD31 antibodies. Samples were analysed for Vegfr2 messenger RNA expression by quantitative reverse transcriptase PCR and data were normalized for CD31 expression. The primers are listed in Supplementary Table 1.

Small interfering RNA b-galactosylceramidase knockdown in human umbilical vein endothelial cells Human umbilical vein endothelial cell silencing was carried out with a pool of lentiviral particles containing three short hairpin RNA targetspecific constructs against human GALC (sc-60669-V; Santa Cruz Biotechnology) whereas short hairpin RNA lentiviral particles encoding scrambled short hairpin RNA sequence (sc-108080; Santa Cruz Biotechnology) were used as controls. Cells were infected for 7 h in medium containing 8 mg/ml of polybrene with 5  104 lentiviral particles. Puromycin (0.8 mg/ml) was added 24 h later as a selection agent.

M. Belleri et al.

Psychosine quantification Lipids were isolated from 4  106 human umbilical cells/sample as described (Galbiati et al., 2007). The tion was analysed using a high performance liquid tandem mass spectrometry (LC/MS/MS) system and centration was normalized to protein concentration.

vein endothelial psychosine fracchromatography psychosine con-

Human umbilical vein endothelial cell motility assay Cell motility was assessed by time lapse videomicroscopy. Briefly, cells seeded on fibronectin-coated dishes were stimulated with 30 ng/ml VEGF and phase-contrast snap photographs (one frame every 10 min) were digitally recorded for 8 h. Cell paths (20–50 cells per experimental point) were generated from centroid positions and migration parameters were analysed with the ‘Chemotaxis and Migration Tool’ of ImageJ software.

Histopathology of human globoid cell leukodystrophy biopsy Autopsy brain specimens of one patient with globoid cell leukodystrophy (aged 2.5 years) and a control brain obtained anonymously from a patient with no neurological abnormalities, who died for respiratory complications, were obtained from the Eunice Kennedy Shriver National Institute of Child Health and Human Development at the University of Maryland School of Medicine. Use of this material was approved by the Institutional Review Board of College of Medicine, University of Illinois. Formalin-fixed, paraffin-embedded tissue sections were haematoxylin and eosin and single or double immunohistochemical stained. The following primary antibody and dilutions were used: polyclonal anti-GFAP (Dako; 1:200), monoclonal anti-CD68 (Dako; clone KP1, 1:300), monoclonal anti-SMA (Dako; clone HHF35, 1:100), monoclonal anti-CD31 (Novocastra; clone 4C9, 1:50), polyclonal anti-Factor VIII-related antigen (Novocastra; 1:100).

Statistical analysis Comparisons among multiple groups were performed by ANOVA. Comparisons between two groups were achieved with Student’s unpaired t-test.

Results Psychosine inhibits the angiogenesis process by acting as a cytoskeleton-disrupting agent In a first set of experiments, we investigated the capacity of psychosine to affect endothelial cells in vitro. Because of endothelial cell heterogeneity (Bastaki et al., 1997; Eberhard et al., 2000; Chi et al., 2003), the anti-angiogenic potential of psychosine was assessed on macrovascular and microvascular endothelial cells of different origin. As shown in Fig. 1A, psychosine inhibits the proliferation of subconfluent murine aortic endothelial cells in a 24 h short-term cell proliferation assay. The effect was dose-dependent, the

GALC deficiency and angiogenesis

Brain 2013: 136; 2859–2875

| 2863

Figure 1 Psychosine is an actin cytoskeleton-destabilizing agent in endothelial cells. (A) Murine aortic endothelial cells were seeded at 20 000 cells/cm2 (for the 24 h and 72 h experiments) or 2500 cells/cm2 (for the 120 h experiment). After overnight incubation (T0), cells were incubated with vehicle or increasing concentrations of psychosine for 24 h (filled circle),72 h (open circle) or 120 h (filled triangle). Then, cells were trypsinized and counted in a Burker chamber. Data are the mean  SEM of three to five determinations. (B and C) Confluent murine aortic endothelial cells were wounded with a 1.0-mm-wide rubber policeman and incubated with vehicle or increasing concentrations of psychosine. (B) Three hours after wounding, cells were immunostained with an anti-pericentrin antibody. The percentage of cells at the edge of the wound with microtubule organization centre (MTOC) localized in the front quadrant was then calculated (n = 70–80). *Random distribution of the microtubule organization centre around the nucleus, as observed in the quiescent monolayer, results in 25% of cells with microtubule organization centre located in the front quadrant. Inset: Representative image of microtubule organization centre positioning in the front quadrant of a migrating cell at the edge of the wound. (C) Sixteen hours after wounding, cells were photographed and the percentage of wound repair was quantified by computerized analysis of the digitalized images. Insets: Representative images of vehicle and of 100 mM psychosine-treated monolayers photographed 16 h after wounding. Data are the mean  SD of three to four determinations. (D) Subconfluent bovine aortic endothelial cells were incubated with vehicle (open bars) or 50 mM psychosine (black bars) and stained with rhodamine-phalloidin at the indicated time points. The percentage of cells (n = 200–270) with actin disassembly was then calculated. Data are the mean  SD of three determinations. Insets: Representative images of rhodamine-phalloidin stained cells treated for 6 h with vehicle or 50 mM psychosine.

concentration causing 50% inhibition (ID50 value) being equal to 60 mM. Of note, the inhibitory potency of psychosine was progressively increased when assessed in a long-term cell proliferation assay in which murine aortic endothelial cells were incubated with psychosine for 72 h or 120 h, its ID50 value decreasing to 15 mM and 5.0 mM, respectively (Fig. 1A). Also, glucopsychosine exerted an inhibition on murine aortic endothelial cell proliferation similar

to psychosine, whereas only a limited inhibitory effect was observed when cells were treated under the same experimental conditions with the negative control N-acetyl-D-sphingosine (Im et al., 2001), thus confirming the specificity of the effect (Supplementary Fig. 1). Inhibition of cell proliferation occurred in the absence of significant apoptotic events in psychosine-treated cells that recovered a normal proliferation rate when incubated in

2864

| Brain 2013: 136; 2859–2875

fresh culture medium after 24 h treatment with the compound. Moreover, no effect on cell number was observed when psychosine was administered to confluent cells, indicating that psychosine was devoid of acute cytotoxic effects on quiescent endothelial cells (data not shown). Similar results were obtained for FGF2-overexpressing FGF2-T-MAE cells in which endogenous FGF2 drives cell proliferation via an autocrine loop of stimulation (Gualandris et al., 1996) and for serum-activated bovine aortic endothelial cells and murine lung microvascular 1G11 endothelial cells (Dong et al., 1997) (data not shown). On this basis, the capacity of psychosine to affect different steps of the angiogenesis process, including endothelial cell migration and extracellular matrix invasion, was investigated. Migration of endothelial cells following the mechanical wounding of a cell monolayer is characterized by the repositioning of the cell microtubule organization centre from a random distribution to a biased localization in front of the nucleus, toward the direction of cell migration (Ueda et al., 1997). Consistently, 3 h after wounding the microtubule organization centre was located at the leading edge in the front of the nucleus in 55% of murine aortic endothelial cells adjacent to the wound. As shown in Fig. 1B, psychosine prevented microtubule organization centre repositioning in a dose-dependent manner, thus impairing the ability of treated cells to polarize toward the migratory stimulus. Accordingly, psychosine treatment caused a significant delay in the healing of wounded endothelial monolayers of murine aortic endothelial and 1G11 cells (Fig. 1C and data not shown). A similar inhibition was exerted by glucopsychosine whereas N-acetyl-D-sphingosine was ineffective (data not shown). The ability of psychosine to inhibit endothelial cell migration was confirmed by its capacity to prevent the formation of cell membrane ruffles at the edge of the wounded 1G11 cell monolayer in a time- and dose-dependent manner (Supplementary Fig. 2A). Also, psychosine inhibited the capacity of bovine aortic endothelial cells to invade a 3D type I collagen gel and to organize capillary-like structures when stimulated by exogenously added FGF2 plus VEGF (Supplementary Fig. 2B) and the capacity of endogenously activated FGF2-T-MAE cells to migrate and invade a fibrin gel (Gualandris et al., 1996), thus inhibiting the formation of radially growing sprouts (Supplementary Fig. 2C). Previous observations had shown the ability of psychosine to affect actin reorganization, leading to the formation of multinuclear globoid cells (Kanazawa et al., 2000; Kozutsumi et al., 2002). On this basis, we evaluated the effect of psychosine on different components of endothelial cell cytoskeleton. As shown in Fig. 1D, treatment with 50 mM psychosine causes the disassembly of actin stress fibres in subconfluent bovine aortic endothelial cells. The effect was time-dependent, 50% of the cells showing actin disassembly 3–6 h after exposure to the compound. No effect was instead exerted by psychosine on endothelial microtubule integrity and intermediate filament organization (data not shown). In keeping with these observations, psychosine caused a dramatic impairment of the capacity of bovine aortic endothelial cells to spread on immobilized fibronectin without affecting their ability to adhere to the substratum (Supplementary Fig. 3). To assess the capacity of psychosine to affect neovascularization in vivo, gelatin sponges adsorbed with FGF2 alone or added with

M. Belleri et al. psychosine were implanted on the top of the chorioallantoic membrane of 8-day-old chick embryos (Ribatti et al., 1997). Sponges adsorbed with vehicle were used as controls. As shown in Fig. 2A and B, psychosine caused a significant inhibition of FGF2-induced angiogenesis in the chorioallantoic membrane. This occurred in the absence of any effect on embryonic development and survival. To further demonstrate that a single high-dose bolus of psychosine is sufficient to inhibit the angiogenic response triggered in vivo by FGF2, MatrigelTM plugs containing PBS or 300 ng FGF2 were injected subcutaneously in the flank of C57BL/6 mice in the presence of 200 mM psychosine dissolved in dimethyl sulphoxide or of

Figure 2 Psychosine inhibits angiogenesis in vivo. (A) Chorioallantoic membranes were implanted at Day 8 with gelatin sponges containing vehicle (open bar) or 500 ng FGF2 in the absence (black bar) or presence (dashed bar) of 1.6 mM psychosine. At Day 12 newly-formed blood vessels converging versus the implant were counted under a stereomicroscope. Data are the mean  SD of eight determinations. *P 5 0.01. (B) Representative images of chorioallantoic membranes treated with FGF2 in the absence (a) or in the presence (b) of psychosine. (C) C57BL6 mice were injected subcutaneously with 400 ml MatrigelTM containing PBS or 300 ng FGF2 in the presence of 200 mM psychosine dissolved in dimethyl sulphoxide or an equal volume of vehicle. After 7 days, the vascular response was quantified by evaluation of the levels of expression of the endothelial markers CD31, Vegfr2 and VE-cadherin by quantitative reverse transcriptase PCR. Relative expression levels were normalized in respect to Gapdh messenger RNA levels and data are the mean  SEM of seven to eight animals per group. *P 5 0.05.

GALC deficiency and angiogenesis

Brain 2013: 136; 2859–2875

| 2865

an equal volume of vehicle. After 7 days, plugs were harvested and the vascular response was quantified by evaluation of the levels of expression of the endothelial markers CD31, Vegf receptor-2 (Vegfr2) and VE-cadherin by quantitative reverse transcriptase PCR. The results demonstrate that psychosine also causes a potent inhibition of the angiogenic activity exerted by FGF2 in the MatrigelTM plug assay, as shown by the significant decrease in the levels of all the endothelial cell markers in psychosine-treated in respect to vehicle-treated FGF2 plugs (Fig. 2C). Taken together, these data demonstrate that psychosine inhibits the angiogenic process by acting on endothelial cells as an actin cytoskeleton-disrupting agent. These findings prompted us to assess the capacity of endothelial cells to express GALC and the effect of GALC deficiency on CNS microvascularity in twitcher mice, an authentic model of globoid cell leukodystrophy (Kobayashi et al., 1980; Suzuki, 1995). Twitcher mice develop most of the pathological hallmarks seen in human globoid cell leukodystrophy patients, including psychosine accumulation (Whitfield et al., 2001).

Defective brain vascularity in twitcher mouse In a preliminary set of experiments we evaluated GALC expression in murine endothelial cells of different origin. GALC activity was detectable in the extract of murine macrovascular and microvascular endothelial cells, including murine aortic endothelial cells, 1G11 cells, brain microvascular endothelial cells (Bastaki et al., 1997) and dermal microvascular murine subcutaneous endothelial cells (Dong et al., 1997), with highest levels in brain endothelium (Supplementary Fig. 4A). Accordingly, all the murine endothelial cell lines tested express Galc gene transcripts (data not shown). In agreement with these observations, immunohistochemical analysis confirmed GALC expression in the murine endothelium in vivo, a strong immunoreactive signal being observed in brain microvessels (Supplementary Fig. 4B). In humans, cerebral microvasculature undergoes considerable angiogenesis in the foetus and premature infant until it reaches a constant density that is maintained throughout adulthood (Mito et al., 1991). Similarly, a significant increase in microvascular density due to capillary neovascularization processes occurs during post-natal development in rodent CNS in which cortical angiogenesis peaks between post-natal Days 13 and 24 and stabilizes between Days 24 and 33 (Ogunshola et al., 2000). On this basis, CNS microvascular density was assessed by immunohistochemical analysis of the brain of wild-type, heterozygous carrier (twi/ + ) and homozygous (homozygous) juvenile mice using FVIII/ vWF as a brain capillary blood vessel marker (DeBault and Cancilla, 1980). The analysis was performed on coronal sections of the telencephalon and microvascular density was measured in the cortex and periventricular areas, both characterized by a robust post-natal angiogenic response (Breier et al., 1992; Ogunshola et al., 2000). Measurements were done at Days 12 and 17 before the onset of evident neurological signs, at Day 24 when pathological alterations are occasionally detectable, and at Day 36 when homozygous mice show clear neurological defects,

Figure 3 Defective brain vascularity in twitcher mouse. (A) Wild type (wt), heterozygous carrier (twi/ + ) and homozygous (twi/ twi) mice were sacrificed at Days 12, 17, 24 and 36 (three to four animals per group). Microvascular density of cerebral cortex (a) and periventricular area (b) were evaluated by quantification of FVIII/vWF + blood vessels. Four sections per sample were analysed with a square reticulum (0.78 mm2) inserted in the eyepiece by counting capillary blood vessel density in four to six 200 fields covering almost the whole of each section by two investigators with a double-headed light microscope. (B) Representative images of cerebral cortex FVIII/vWF + capillary blood vessels in Day 36 homozygous brain (a) when compared with wild-type tissue (b). Scale bar = 50 mm. (C) Sections (50 mm) of Day 36 (P36) wild-type (a) and homozygous (b) brains were decorated with an anti-CD31 antibody, Z-stack images of frontal cortex were acquired with an epifluorescence microscope equipped with ApoTome optical sectioning device, image processing was carried out with AxioVision 4.8 software and 3D renderings were obtained using ImageJ software. Scale bar = 50 mm.

including tremor and hind limb paralysis. As shown in Fig. 3A, cortical density of FVIII/vWF + capillary blood vessels increases from Day 12 to Day 24 in both wild-type and heterozygous animals to remain constant thereafter. At variance, no increase of cortical microvascular density was observed in homozygous animals whose brain vascularization remained similar to that measured in Day 12 control mice throughout the whole experimental period and was significantly reduced (P 5 0.05 or better) when compared with that observed in Day 17–36 wild-type and

2866

| Brain 2013: 136; 2859–2875

heterozygous mice (Fig. 3A and B). Accordingly, a reduced capillary vascularity was observed using CD31 as endothelial marker in 3D reconstructions of cortex sections from Day 36 homozygous mice when compared with control mice (Fig. 3C). A significant reduction in microvascular density was observed also in the periventricular area of homzygous mice brains, indicating that GALC deficiency affects microvascularization of both the grey and white matter in the CNS of twitcher mice (Fig. 3A). No changes in the levels of expression of the major proangiogenic factors VEGF-A, FGF2 and angiopoietin 2 (Angpt 2) were observed in the brain of homozygous mice until Day 24 when compared with wild-type and heterozygous animals (Supplementary Fig. 5). At Day 36, a limited decrease of VEGF messenger RNA levels was paralleled by a dramatic upregulation of FGF2 expression whereas the levels of Angpt transcripts remained unchanged. Thus, the observed defects in homozygous brain microvascularity do not appear to be the direct consequence of a decreased expression of pro-angiogenic factors. In parallel with the defects in microvascular density, capillaries with dilated morphology were frequently found in homozygous brains when compared with controls (Supplementary Fig. 6). Also, ultrastructural analysis of the frontal cortex microvasculature showed significant alterations of microvascular endothelium in the CNS of homozygous mice. Day 36 homozygous vessels are lined by an irregular endothelium with luminal projections, enveloped by a thick basement membrane. As observed in the brain of dystrophin-deficient mice (Nico et al., 2003), homozygous vessels were characterized by endothelial cells surrounded by swollen glial endfeet and sealed by altered tight junctions with detached plasma membrane leaflets (Fig. 4A). Accordingly, confocal microscopy analysis of the homozygous cortical microvasculature showed a remarkable decrease of the expression of ZO-1, a tight junctionassociated protein functionally critical in regulating endothelial barrier integrity (Lee et al., 2006), with a discontinuous staining when compared with controls (Fig. 4B). Accordingly, intracardiac injection of sulpho-NHS-LC-biotin followed by Alexa FluorÕ 488conjugated streptavidin staining of brain sections showed an increase of brain vessel permeability with extravascular leakage of the injected tracer in both white and grey matter areas of Day 36 homozygous (twi/twi) mice (Fig. 4C).

Defective angiogenic responses in twitcher mouse The data above indicate that GALC deficiency affects the ability of twitcher endothelium to respond to post-natal pro-angiogenic stimuli. On this basis, to assess the capacity of twitcher mice to mount an angiogenic response when stimulated by exogenously administered pro-angiogenic factors, homozygous animals were injected subcutaneously with a MatrigelTM plug containing the angiogenic factors FGF2 plus VEGF in the post-natal periods Day 12, Days 17–24 and Days 29–36. After 7 days, plugs were harvested and neovascularization was quantified by computerized image analysis of the CD31 + vessels invading the implant. As shown in Fig. 5A–C, a progressively reduced angiogenic response was observed in homozygous mice when compared with wild-type

M. Belleri et al. and heterozygous carrier animals. A similar neovascularization was observed in the plugs of the three experimental groups when the animals were implanted at Day 12 whereas a significant reduction of plug vascularization occurred in homozygous mice at Days 17–24 and Days 29–36. Quantitative reverse transcriptase PCR analysis of the levels of CD31 messenger RNA in MatrigelTM plugs confirmed the paucity of the neovascular response in Day 29–36 homozygous mice when compared with wild-type and heterozygous carrier animals (Fig. 5D). To further assess the capacity of homozygous endothelium to respond to an angiogenic stimulus, thoracic aortas were isolated from wild-type, heterozygous carrier and homozygous mice at Days 12, 24 and 36. Aorta rings were embedded in 3D-fibrin gel and vessel sprouts invading the gel were counted under a stereomicroscope 6 days after incubation with 30 ng/ml VEGF. The capacity to sprout in response to VEGF was lost in aorta rings isolated from Day 24 and Day 36 homozygous animals whereas it was fully retained in aorta rings isolated from Day 12 homozygous mice (Fig. 6A and B). This lack of response occurred despite the similar levels of expression of VEGFR2, the main transducer of VEGF-mediated angiogenic signals (Olsson et al., 2006), in all the aorta rings examined (Fig. 6C). Also, histological and CD31 immunohistochemical analyses did not show any apparent morphological alteration in homozygous aorta rings that were lined by a regular CD31 + endothelium (Fig. 6D and E). To confirm that the lack of response of homozygous endothelium was due to the deficit in GALC activity, Day 24 homozygous aorta rings were infected 24 h before VEGF stimulation with a lentivirus harbouring the murine Galc complementary DNA. As shown in Fig. 6F, Galc lentiviral infection was able to rescue the capacity of homozygous aorta rings to sprout in response to VEGF stimulation whereas infection with a control EGFP-harbouring lentivirus was ineffective.

b-Galactosylceramidase deficiency affects human endothelium In order to assess the impact of GALC downregulation in human endothelium, human umbilical vein endothelial cells were infected with a pool of concentrated lentiviral particles containing short hairpin RNA target-specific constructs designed to knockdown the human GALC gene. After puromycin selection, infected human umbilical vein endothelial cells were maintained under standard cell culture conditions and assessed for the levels of GALC activity, psychosine concentration, proliferative capacity and motogenic activity. Silenced cells showed a significant decrease in GALC activity when compared with uninfected cells or with cells infected with a scrambled control short hairpin RNA (Fig. 7A). This was paralleled by a significant change of the levels of psychosine in the cell extracts of GALC-silenced cells that increased from 7.2–10.9 pmoles/mg of protein, as measured in uninfected cells or cells infected with the scrambled control short hairpin RNA, to 16.1 pmoles/mg of protein in GALC short hairpin RNA infected human umbilical vein endothelial cells. Also, GALC downregulation resulted in a significant decrease in the rate of growth of

GALC deficiency and angiogenesis

Brain 2013: 136; 2859–2875

| 2867

Figure 4 Structural analysis of brain microvasculature in twitcher mouse. (A) Transmission electron microscopy of frontal cortex microvasculature in the brain of Day 36 homozygous (a–c) and wild-type (d) mice. (a) Homozygous microvessels are lined by an irregular endothelium with luminal projections (arrow), enveloped by a thick basement membrane (arrowhead). (b and c) Homozygous microvessels show endothelial cells sealed by altered tight junctions (arrows) with detached plasma membrane leaflets (arrow in c) and surrounded by swollen glial endfeet (asterisk in b). (d) Control wild-type microvessels with regular endothelial cells connected by normal tight junctions (arrow) and surrounded by flat glial endfeet containing glycogen granules (asterisk). Scale bars: a, b and d = 0.6 mm; c = 0.2 mm. (B) Confocal immunofluorescence analysis of ZO-1 + endothelial junctions in the brain of Day 36 wild-type (a) and homozygous (b) mice. Frontal cortex sections were double-immunostained with anti-ZO-1 (in red) and anti-GFAP (in green) antibodies and analysed under a confocal laser scanning microscope. Control wild-type vessels (a) are surrounded by GFAP + labelled glial endfeet (arrow) and show an intense ZO-1 + red fluorescence signal along endothelial cell-cell contacts (arrowheads), almost absent in homozygous brain sections (b). Images are representative of two independent experiments in which three individual mice were analysed. Scale bar = 15.8 mm. (C) Vessel permeability was analysed by intracardiac injection of sulpho-NHS-LC-biotin in Day 36 wild-type (a) and homozygous (b) mice. Sections (100 mm) of biotinylated brains were stained with Alexa Fluor 488-conjugated streptavidin and Z-stack images were acquired with an epifluorescence microscope equipped with ApoTome optical sectioning device. Scale bar = 200 mm. Sulpho-NHS-LC-biotin diffuses through the endothelial barrier and accumulates in the extravascular environment in the frontal cortex of homozygous brain. Identical results were obtained in white matter areas (not shown). wt = wild-type; twi/ + = heterozygous carrier; twi/twi = and homozygous.

GALC-silenced cells maintained in growth factor-supplemented cell culture medium (Fig. 7B) and by their incapacity to proliferate in response to VEGF when maintained under low-serum conditions (Fig. 7C). Finally, silenced human umbilical vein endothelial cells showed a reduced motility in response to VEGF stimulation as

assessed by time-lapse videomicroscopy (Fig. 7D). Of note, similar levels of VEGFR2 phosphorylation were observed in GALCsilenced and in control cells following VEGF stimulation, indicating that downregulation of GALC activity does not directly affect VEGFR2 activation (data not shown). In conclusion, similar to

2868

| Brain 2013: 136; 2859–2875

M. Belleri et al.

Figure 5 Defective in vivo angiogenic response in twitcher mouse. MatrigelTM plugs containing PBS or 500 ng FGF2 plus 100 ng VEGF were implanted subcutaneously in wild-type (wt), heterozygous carrier (twi/ + ) and homozygous (twi/twi) mice at Days 12, 17–24 or 29–36 days after birth (7–12 animals per group). After 7 days, plugs were harvested [representative Day 29–36 plugs are shown in (A)] and frozen sections were immunostained with an anti-CD31 antibody (B). The vascular response was quantified by computerized image analysis of CD31 + areas (C) and by quantitative reverse transcriptase PCR analysis of the levels of expression of the endothelial marker CD31 normalized for Gapdh expression in Day 29–36 plugs (D). *Statistically different from wild-type and heterozygous animals (P 5 0.05 or better). Scale bar = 60 mm.

murine endothelium, human endothelial cells express GALC activity whose deficiency impairs their ability to respond to angiogenic stimuli. Due to the difficulties to obtain bioptic material, only scarce pathological studies about brain angioarchitecture have been reported on a very limited number of leukodystrophic patients, including one 8-month-old boy with globoid cell leukodystrophy (Kondo and Suzuki, 1993; Buee et al., 1994). On this basis, we have analysed the microvasculature of brain biopsies obtained from a 2.5-year-old patient with globoid cell leukodystrophy and an age-matched control. The globoid cell leukodystrophy brain biopsy showed key neuropathological features of an advanced disease stage with extensive demyelination and severe loss of oligodendrocytes, axonal degeneration and diffuse reactive gliosis

associated with the presence of CD68 + multinucleated globoid cells peculiarly distributed around small blood vessels (Fig. 8A). As anticipated, the white matter was found to be predominantly affected, even though activated microglial cells and mild gliosis were found also in grey matter and periventricular regions. Even though the advanced stage of the disease did not allow appreciating in detail the vessel morphology and distribution in the white matter of the globoid cell leukodystrophy specimen, we observed a marked decrease of microvascular density in cortical regions and periventricular areas with vessels characterized by a larger dilated lumen compared with the control (Fig. 8B). Also, double immunostaining of the grey matter from the globoid cell leukodystrophy patient specimen highlighted numerous reactive GFAP + astrocytes surrounding blood vessels lined by numerous

GALC deficiency and angiogenesis

Brain 2013: 136; 2859–2875

| 2869

Figure 6 Defective ex vivo angiogenic response of twitcher endothelium. Aorta rings were isolated from wild-type (wt), heterozygous carrier (twi/ + ) and homozygous (twi/twi) mice at Days 12, 24 or 36 (10–20 aorta rings per group) and embedded in 3D-fibrin gel in the absence (PBS) or in the presence of 30 ng/ml VEGF. After 6 days, aorta rings were photographed (representative aorta rings from Day 36 mice are shown in A) and vessel sprouts were counted under a stereomicroscope (B). *Statistically different from wild-type and heterozygous animals (P 5 0.01). Aortae (n = 7) were isolated from Day 24–36 mice and analysed for the levels of VEGFR2 messenger RNA expression by quantitative reverse transcriptase PCR and data were normalized for CD31 expression (C). Histological sections of aortae from the same animals were stained with haematoxylin and eosin (H&E) (D) or were decorated with anti-CD31 antibodies (E). No apparent morphological alteration was observed in homozygous aorta rings that were lined by a regular CD31 + endothelium (arrowheads in E). Images are representative of two independent experiments in which three individual mice were analysed. (F) Aorta rings isolated from Day 24 homozygous mice were incubated in serum-free medium with 8 mg/ml polybrene and lentiviral particles harbouring the murine GALC or control EGFP complementary DNA (cDNA). Mock-infected wild-type aorta rings were used as controls. After 24 h, rings were embedded in fibrin gel and incubated with 30 ng/ml VEGF in the presence of 10% foetal calf serum. Then, vessel sprouts were counted under a stereomicroscope at 5 days. Data are the mean  SD of six to eight rings from two animals per group. *P 5 0.01. Scale bars: D = 100 mm; E = 15 mm.

glial endfeet projections and irregular shaped factor VIII + endothelium with discontinuous wrapping by smooth muscle actin (SMA) + cells (Fig. 8B). Thus, in keeping with the data obtained in twitcher mice, our observations indicate that GALC deficiency may affect human blood vessel endothelium in patients with globoid cell leukodystrophy.

Discussion In this study, psychosine has been identified as an endothelial actin disassembling agent endowed with anti-angiogenic activity in vitro and in vivo. Psychosine inhibits different steps of the angiogenesis process in vitro, including endothelial cell proliferation, migration,

2870

| Brain 2013: 136; 2859–2875

M. Belleri et al.

Figure 7 Small interfering RNA GALC knockdown inhibits human umbilical vein endothelial cell proliferation and motility. Human umbilical vein endothelial cells were infected with lentiviral particles harbouring GALC-targeting or scrambled (scr) short hairpin RNAs (shRNA). After puromycin selection, GALC-shRNA (black bar) and scr-shRNA (dashed bar) infected cells were compared with control uninfected cells (open bar) for GALC enzymatic activity using a fluorometric assay (A). Also, control uninfected (open triangle), scr-shRNA (open circle) and GALC-shRNA (filled circle) infected cells were maintained in complete growth factor-supplemented cell culture medium and counted in a Burker chamber at the indicated time points (B). Control uninfected, scr-shRNA and GALC-shRNA infected cells were treated with vehicle (black bars) or with 30 ng/ml VEGF (open bars) and assessed for their capacity to respond to the growth factor under low serum conditions in a 72 h cell proliferation assay (C) and in a time lapse videomicroscopy cell migration assay (D). Proliferation data are the mean  SD of three determinations and motility data are the mean  SD of 50 cells per experimental point. *Statistically different from vehicle (P 5 0.01).

extracellular matrix invasion and sprouting, and hampers the activity exerted in vivo by the potent pro-angiogenic factor FGF2 in the chick embryo chorioallantoic membrane and murine MatrigelTM plug assays. Previous observations had shown the ability of psychosine to affect actin reorganization, leading to inhibition of cytokinesis and formation of microglia and macrophage-derived multinuclear globoid cells, a marker of globoid cell leukodystrophy (Kanazawa et al., 2000; Kozutsumi et al., 2002). Our data demonstrate that psychosine causes a rapid disorganization of actin stress fibres in endothelial cells and hampers their ability to spread on

immobilized fibronectin with no effect on their adhesive capacity to the substratum. In keeping with recent observations about the involvement of actin flow in microtubule organization centre re-orientation and polarized cell migration (Gomes et al., 2005; Ang et al., 2010), actin cytoskeleton disruption was paralleled by the incapacity of psychosine-treated endothelial cells to orient their microtubule organization centre at the leading front of a wounded monolayer and to form cell membrane ruffles. Centrosome repositioning stabilizes the direction of migration (Ueda et al., 1997). Accordingly, the incapacity of psychosinetreated cells to re-orient their microtubule organization centre

GALC deficiency and angiogenesis

Brain 2013: 136; 2859–2875

| 2871

Figure 8 Histopathology and vessel alterations in brain biopsy from human globoid cell leukodystrophy (GLD) disease. (A) The brain biopsy from a 2.5-year-old patient with globoid cell leukodystrophy shows features of an advanced disease stage (a), including intense and diffuse astrogliosis highlighted by GFAP + immunostaining (b) and the presence of numerous CD68 + multinucleated globoid cells peculiarly distributed around blood vessels (c). (B) Representative images from cortical regions of globoid cell leukodystrophy (a–c) and control (d–f) brain biopsies highlighting alterations of the cerebral vascularization. In contrast with the control, the globoid cell leukodystrophy biopsy shows a marked decrease of FVIII + microvascular density (a versus d), numerous reactive GFAP + astrocytes surrounding FVIII + blood vessels and projecting numerous glial endfeet around an irregular shaped endothelium (b versus e) and partial loss of SMA + vascular smooth muscle cell coverage (c versus f). All images are from  20 original magnification (scale bar = 100 mm) except for panels c, f and insets in panels b, e that are from 40 original magnification (scale bar = 50 mm).

and to stabilize pseudopod extensions results in impaired cell migration and incapacity to heal the wounded endothelial cell monolayer. In keeping with the ability of cytoskeleton-destabilizing agents to affect proliferating endothelial cells but not the quiescent endothelium (Dark et al., 1997; Iyer et al., 1998), psychosine did not exert a significant effect on contact-inhibited confluent endothelial cells. The cytostatic and actin cytoskeleton disassembly activity of psychosine on activated endothelium also explains the ability of this lysolipid metabolite to inhibit the capacity of endothelial cells to invade 3D extracellular matrices. These effects result in a significant inhibition of the neovascular response triggered in vivo by FGF2. It must be pointed out that psychosine has been shown to inhibit protein kinase C and AMP activated kinase activities, to interfere with insulin-like growth factor 1 signalling, and to activate phospholipase A2 (reviewed in Ballabio and Gieselmann, 2009). Together with the observed effects on actin cytoskeleton assembly, alterations in intracellular signal transduction pathway(s) may also contribute to the angiostatic properties of psychosine.

Short-term versus long-term cell proliferation assays demonstrate that the angiostatic potency of psychosine increases with time of exposure of endothelial cells to the compound. A decrease in the ID50 value from 60 mM to 5.0 mM was observed when cells were exposed to psychosine for 120 h rather than for 24 h. Thus, long-lasting exposure to psychosine dramatically increases its capacity to affect endothelial cell functions at concentrations that are compatible with the micromolar levels of this metabolite measured in the CNS and in non-nervous tissues/organs of patients with globoid cell leukodystrophy and/or of twitcher mice (see below). This suggests that the progressive, chronic accumulation of this GALC metabolite may cause a significant effect on endothelium in globoid cell leukodystrophy. Accordingly, our data demonstrate a significant decrease of microvascular density and alterations of angio-architecture of capillary blood vessels in the post-natal brain of juvenile twitcher mice, an authentic model of globoid cell leukodystrophy that develops most of the pathological hallmarks seen in human patients, including psychosine accumulation (Kobayashi et al., 1980; Suzuki, 1995). Capillary defects were observed in the

2872

| Brain 2013: 136; 2859–2875

brain of twitcher mice at Days 17–24 before disease onset, as well as at Day 36 during disease progression whereas no vascular defects were observed at Day 12. Also, the capacity of homozygous animals to respond to angiogenic stimuli in the MatrigelTM plug and aorta ring assays was lost at Days 17–36 whereas it was fully retained at Day 12. These data are in keeping with the normal embryonic and early post-natal development of twitcher mice and support the hypothesis that the progressive accumulation of psychosine consequent to GALC deficiency is responsible for the incapacity of post-natal homozygous endothelial cells to respond to pro-angiogenic stimuli. Enzyme knockdown by small interfering GALC targeting short hairpin RNA results in the intracellular accumulation of psychosine in human umbilical vein endothelial cells paralleled by a significant decrease in the mitogenic and motogenic response to VEGF. This occurred regardless of the fact that GALC short hairpin RNA transduction was able to induce only a partial decrease of GALC activity in human umbilical vein endothelial cells, in contrast with the lack of vascular alterations we observed in the endothelium of heterozygous twi/ + mice. Despite significant differences between the two experimental models that may be responsible for this apparent discrepancy (e.g. in vitro versus in vivo conditions with possible adaptive responses that may occur during embryonic development but absent in silenced cells), the data demonstrate that GALC deficiency hampers the capacity of endothelial cells to respond to angiogenic stimuli. Relevant to this point, the capacity of homozygous aorta rings to sprout following VEGF stimulation was rescued after gene transfer using lentiviral vectors carrying the GALC complementary DNA. However, either single or combined therapies based on the intravenous administration of congenic bone marrow cells and GALC-lentiviral vectors in twitcher pups before Day 2 (Galbiati et al., 2009) were unable to rescue the angio-architecture changes occurring in the brain of homozygous mice (data not shown). Indeed, therapies such as transplantation of haematogenous cells, neural stem cells, enzyme replacement and gene therapy have all showed to be very limited in preventing the development of the disease in twitcher mice (Galbiati et al., 2009). Similar limitations have been described in human patients undergoing bone marrow transplantation (Wenger et al., 2000; Sakai, 2009). While the reasons for these limitations are still not clear, timing and extension of the enzymatic correction are at the core of the problem. Indeed, correction of lysosomal storage diseases such as globoid cell leukodystrophy are largely based upon the concept of enzyme cross-correction in which lysosomal enzymes may follow the secretory pathway and be taken up by neighbouring cells through a receptor-mediated process (Kolter and Sandhoff, 2006). GALC sources (i.e. donor-derived macrophages) are known to take several weeks to infiltrate the nervous system of the recipient, before cross-correction in the brain, spinal cord and nerves can take place (Galbiati et al., 2009). This leaves various pathological aspects of the disease essentially untreated for a significant period of post-natal development. Our work shows changes of angioarchitecture in Day 17 twitcher mice, suggesting that vasculature deficits were already established by the moment of GALC correction (Galbiati et al., 2009).

M. Belleri et al. So far, the clinical and pathological manifestations of globoid cell leukodystrophy have been almost exclusively restricted to the nervous system as a consequence of high levels of psychosine accumulation in white matter and peripheral nerves leading to a progressive neurodegeneration. Indeed, the levels of psychosine in the brain of patients with globoid cell leukodystrophy have been reported to vary between 1.0 and 43 nanomoles/g of wet tissue (corresponding to a 1.0–43 mM range of concentrations) (Svennerholm et al., 1980; Igisu and Suzuki, 1984; Li et al., 2011) and similar values have been obtained for the brain of twitcher mice (Igisu and Suzuki, 1984; Nozawa et al., 1992; Whitfield et al., 2001; Contreras et al., 2008; White et al., 2009). Moreover, sensitive assay methods have shown that GALC deficiency causes psychosine accumulation also in nonnervous tissues/organs of twitcher mice, including liver, kidney, bone and lungs among others, with psychosine levels ranging between 0.2 and 3.3 mM (Ida et al., 1982; Kobayashi et al., 1987; Whitfield et al., 2001; Contreras et al., 2008, 2010). Also, nanomolar amounts of psychosine have been measured in the blood of patients with globoid cell leukodystrophy (Chuang et al., 2013) and twitcher animals (Zhu et al., 2012). Accordingly, globoid cell leukodystrophy deficiency in twitcher mice has been shown to be responsible for post-natal bone growth retardation (Contreras et al., 2010), liver damage (Contreras et al., 2008), lymphoid organ atrophy (Galbiati et al., 2007), ultrastructural alterations of the kidney (Takahashi et al., 1984), and impairment of the haematopoietic stem cell niche (Visigalli et al., 2010), different cell types showing a different response to psychosine toxicity (Kanazawa et al., 2000). Our data demonstrate that endothelial cells of different origins express GALC in vitro and in vivo. Accordingly, as stated above, GALC deficiency in twitcher mice and in silenced human umbilical vein endothelial cells affects the capacity of peripheral endothelium to respond to pro-angiogenic factors. Thus, GALC loss-of-function may affect the microvasculature of different organs besides the nervous system by causing the gradual and prolonged accumulation of psychosine to low micromolar tissue concentrations, sufficient to inhibit endothelial cell proliferation after a long-lasting exposure to the molecule. In keeping with this hypothesis, preliminary observations from our laboratory indicate the presence of significant vascular permeability defects in different visceral organs of homozygous mice, including kidney, lung and liver (Supplementary Fig. 7). These observations support the notion that globoid cell leukodystrophy is a generalized psychosine storage disease (Kobayashi et al., 1987; Contreras et al., 2010). Also, they call for future detailed studies about the microvascular architecture of non-nervous tissues/organs in twitcher mice in order to assess the impact of endothelial cell alterations on tissue functionality. Clearly, our results do not rule out the possibility that other defects in sphingolipid metabolisms, besides psychosine accumulation, may concur to the observed alterations of microvasculature following GALC loss-of-function. For instance, twitcher mice show a significant increase in the brain levels of the GALC substrate lactosylceramide (Tominaga et al., 2004), a lipid raft component implicated in cell–cell and cell–matrix interactions and in signalling events linked to cell differentiation, development, apoptosis, and oncogenesis (Chatterjee and Pandey, 2008). Further studies will be

GALC deficiency and angiogenesis required to assess the impact of GALC downregulation on the sphingolipid profile in endothelium and to fully elucidate the molecular mechanism(s) responsible for the vascular defects that follow the loss of GALC activity. A limited number of pathological studies indicate that alterations of the angio-architecture of brain microvasculature and of blood– brain barrier integrity may occur in human leukodystrophies (Kondo and Suzuki, 1993; Buee et al., 1994). In particular, one ultrastructural study has shown some hydropic changes in astrocyte processes around capillaries and a mild enlargement of perivascular space paralleled by macrophage infiltration in one brain biopsy from an 8-month-old patient with globoid cell leukodystrophy (Kondo and Suzuki, 1993). Here, we observed significant defects of the cerebral vascularization of a 2.5-year-old patient with globoid cell leukodystrophy that was characterized by a marked decrease of microvascular density of the cortical region with an irregular shaped endothelium surrounded by numerous reactive GFAP + astrocytes with partial loss of the vascular smooth muscle cell coverage. Overall, human data are in keeping with the vascular alterations observed in the twitcher brain and support the hypothesis that GALC deficiency affects CNS vascularization in human globoid cell leukodystrophy disease. Even though angiogenesis represents a pivotal physiological process during development, reproduction, and wound healing, pathological neovascularization has been shown to be involved in many angiogenesis-dependent diseases, such as cancer, macular degeneration, atherosclerosis, and arthritis (Carmeliet and Jain, 2011). Also, an excess of angiogenic response driven by neuroinflammatory stimuli may contribute to the progression of neurodegenerative disorders, as suggested for the hypervascularization that occurs following accumulation of high levels of amyloid-b in the brain during Alzheimer’s disease (Cameron et al., 2012) and for spinal cord angiogenesis in multiple sclerosis and experimental autoimmune encephalomyelitis (Karlik et al., 2012; MacMillan et al., 2012). Thus, defects in the neurovascular cross-talk due to insufficient or excessive neovascularization may both contribute to the pathogenesis of different disorders of the peripheral and CNS with distinct therapeutic implications (Segura et al., 2009). Although it remains challenging to establish whether vascular alterations in globoid cell leukodystrophy represent the outcome of the astrocytic/neuronal injury, with a consequent deficiency in the production of trophic angioneurins (Lee et al., 2009), or are due to a direct damage of endothelial cell functions, our data indicate that GALC deficiency and psychosine accumulation in the CNS of patients with globoid cell leukodystrophy may directly affect not only the glial/neuronal compartment of the neurovascular brain unit but also its vascular moiety at early stages of postnatal CNS development. A tight cross-talk exists about neurogenesis, and angiogenesis and neovascularization play an important role in post-natal neuroprotection (Ment et al., 1997; Greenberg and Jin, 2005; Zacchigna et al., 2008; Segura et al., 2009). Consequently, alterations in blood vessel development may contribute greatly to the CNS damage occurring during early infancy in globoid cell leukodystrophy. Also, our results reveal a new neuropathogenic aspect of globoid cell leukodystrophy in which the deficiency of GALC activity affects the ability of the blood microvasculature to respond to angiogenic stimuli not only in

Brain 2013: 136; 2859–2875

| 2873

the nervous system but also in somatic organs. This may contribute to worsening globoid cell leukodystrophy disease evolution and may adversely affect therapeutic interventions, including bone marrow repopulation following haematopoietic stem cell transplantation (Visigalli et al., 2010). A better understanding of the role of the vascular system in globoid cell leukodystrophy may point the way towards new angioneurin-based molecular and cellular therapies (Zacchigna et al., 2008).

Acknowledgements We wish to thank Francesca Pagani and Sara Rezzola for helpful technical assistance, Marco Righi (Department of Medical Biotechnology and Translational Medicine, University of Milano) for the precious support in 3D vasculature renderings and Alessandra Biffi (San Raffele Scientific Institute, Milan) for the murine Galc lentiviral construct.

Funding This work was supported in part by grants from Ministero dell’Istruzione, Universita` e Ricerca (MIUR, Centro IDET, FIRB project RBAP11H2R9 2011) and Associazione Italiana per la Ricerca sul Cancro (AIRC grant n 10396) to MP and the National Institutes of Health (NINDS, R01 RNS065808A) to E.R.B. A.G. was supported by a FIRC (Fondazione Italiana per la Ricerca sul Cancro) Fellowship.

Supplementary material Supplementary material is available at Brain online.

References Ang SF, Zhao ZS, Lim L, Manser E. DAAM1 is a formin required for centrosome re-orientation during cell migration. PloS One 2010; 5: 1–10. Ballabio A, Gieselmann V. Lysosomal disorders: from storage to cellular damage. Biochim Biophys Acta 2009; 1793: 684–96. Bastaki M, Nelli EE, Dell’Era P, Rusnati M, Molinari-Tosatti MP, Parolini S, et al. Basic fibroblast growth factor-induced angiogenic phenotype in mouse endothelium. A study of aortic and microvascular endothelial cell lines. Arterioscler Thromb Vasc Biol 1997; 17: 454–64. Belleri M, Ribatti D, Nicoli S, Cotelli F, Forti L, Vannini V, et al. Antiangiogenic and vascular-targeting activity of the microtubuledestabilizing trans-resveratrol derivative 3,5,4’-trimethoxystilbene. Mol Pharmacol 2005; 67: 1451–9. Breier G, Albrecht U, Sterrer S, Risau W. Expression of vascular endothelial growth factor during embryonic angiogenesis and endothelial cell differentiation. Development 1992; 114: 521–32. Buee L, Hof PR, Bouras C, Delacourte A, Perl DP, Morrison JH, et al. Pathological alterations of the cerebral microvasculature in Alzheimer’s disease and related dementing disorders. Acta Neuropathol 1994; 87: 469–80. Cameron DJ, Galvin C, Alkam T, Sidhu H, Ellison J, Luna S, et al. Alzheimer’s-related peptide amyloid-beta plays a conserved role in angiogenesis. PloS One 2012; 7: e39598.

2874

| Brain 2013: 136; 2859–2875

Cantuti-Castelvetri L, Zhu H, Givogri MI, Chidavaenzi RL, LopezRosas A, Bongarzone ER. Psychosine induces the dephosphorylation of neurofilaments by deregulation of PP1 and PP2A phosphatases. Neurobiol Dis 2012; 46: 325–35. Carmeliet P, Jain RK. Molecular mechanisms and clinical applications of angiogenesis. Nature 2011; 473: 298–307. Castelvetri LC, Givogri MI, Zhu H, Smith B, Lopez-Rosas A, Qiu X, et al. Axonopathy is a compounding factor in the pathogenesis of Krabbe disease. Acta Neuropathol 2011; 122: 35–48. Chatterjee S, Pandey A. The Yin and Yang of lactosylceramide metabolism: implications in cell function. Biochim Biophys Acta 2008; 1780: 370–82. Chi JT, Chang HY, Haraldsen G, Jahnsen FL, Troyanskaya OG, Chang DS, et al. Endothelial cell diversity revealed by global expression profiling. Proc Natl Acad Sci USA 2003; 100: 10623–8. Chuang WL, Pacheco J, Zhang XK, Martin MM, Biski CK, Keutzer JM, et al. Determination of psychosine concentration in dried blood spots from newborns that were identified via newborn screening to be at risk for Krabbe disease. Clin Chim Acta 2013; 419: 73–6. Coltrini D, Di Salle E, Ronca R, Belleri M, Testini C, Presta M. Matrigel plug assay: evaluation of the angiogenic response by reverse transcription-quantitative PCR. Angiogenesis 2013; 16: 469–77. Contreras MA, Haq E, Uto T, Singh I, Singh AK. Psychosine-induced alterations in peroxisomes of twitcher mouse liver. Arch Biochem Biophys 2008; 477: 211–8. Contreras MA, Ries WL, Shanmugarajan S, Arboleda G, Singh I, Singh AK. Factors that affect postnatal bone growth retardation in the twitcher murine model of Krabbe disease. Biochim Biophys Acta 2010; 1802: 601–8. Dark GG, Hill SA, Prise VE, Tozer GM, Pettit GR, Chaplin DJ. Combretastatin A-4, an agent that displays potent and selective toxicity toward tumor vasculature. Cancer Res 1997; 57: 1829–34. DeBault LE, Cancilla PA. Some properties of isolated endothelial cells in culture. Adv Exp Med Biol 1980; 131: 69–78. Dong QG, Bernasconi S, Lostaglio S, De Calmanovici RW, MartinPadura I, Breviario F, et al. A general strategy for isolation of endothelial cells from murine tissues. Characterization of two endothelial cell lines from the murine lung and subcutaneous sponge implants. Arterioscler Thromb Vasc Biol 1997; 17: 1599–604. Eberhard A, Kahlert S, Goede V, Hemmerlein B, Plate KH, Augustin HG. Heterogeneity of angiogenesis and blood vessel maturation in human tumors: implications for antiangiogenic tumor therapies. Cancer Res 2000; 60: 1388–93. Galbiati F, Basso V, Cantuti L, Givogri MI, Lopez-Rosas A, Perez N, et al. Autonomic denervation of lymphoid organs leads to epigenetic immune atrophy in a mouse model of Krabbe disease. J Neurosci 2007; 27: 13730–8. Galbiati F, Givogri MI, Cantuti L, Rosas AL, Cao H, van Breemen R, et al. Combined hematopoietic and lentiviral gene-transfer therapies in newborn Twitcher mice reveal contemporaneous neurodegeneration and demyelination in Krabbe disease. J Neurosci Res 2009; 87: 1748–59. Gomes ER, Jani S, Gundersen GG. Nuclear movement regulated by Cdc42, MRCK, myosin, and actin flow establishes MTOC polarization in migrating cells. Cell 2005; 121: 451–63. Greenberg DA, Jin K. From angiogenesis to neuropathology. Nature 2005; 438: 954–9. Gualandris A, Rusnati M, Belleri M, Nelli EE, Bastaki M, MolinariTosatti MP, et al. Basic fibroblast growth factor overexpression in endothelial cells: an autocrine mechanism for angiogenesis and angioproliferative diseases. Cell Growth Differ 1996; 7: 147–60. Haq E, Giri S, Singh I, Singh AK. Molecular mechanism of psychosineinduced cell death in human oligodendrocyte cell line. J Neurochem 2003; 86: 1428–40. Ida H, Umezawa F, Kasai E, Eto Y, Maekawa K. An accumulation of galactocerebroside in kidney from mouse globoid cell leukodystrophy (twitcher). Biochem Biophys Res Commun 1982; 109: 634–8. Igisu H, Suzuki K. Progressive accumulation of toxic metabolite in a genetic leukodystrophy. Science 1984; 224: 753–5.

M. Belleri et al. Ijichi K, Brown GD, Moore CS, Lee JP, Winokur PN, Pagarigan R, et al. MMP-3 mediates psychosine-induced globoid cell formation: implications for leukodystrophy pathology. Glia 2013; 61: 765–77. Im DS, Heise CE, Nguyen T, O’Dowd BF, Lynch KR. Identification of a molecular target of psychosine and its role in globoid cell formation. J Cell Biol 2001; 153: 429–34. Iyer S, Chaplin DJ, Rosenthal DS, Boulares AH, Li LY, Smulson ME. Induction of apoptosis in proliferating human endothelial cells by the tumor-specific antiangiogenesis agent combretastatin A-4. Cancer Res 1998; 58: 4510–4. Jatana M, Giri S, Singh AK. Apoptotic positive cells in Krabbe brain and induction of apoptosis in rat C6 glial cells by psychosine. Neurosci Lett 2002; 330: 183–7. Kanazawa T, Nakamura S, Momoi M, Yamaji T, Takematsu H, Yano H, et al. Inhibition of cytokinesis by a lipid metabolite, psychosine. J Cell Biol 2000; 149: 943–50. Karlik SJ, Roscoe WA, Patinote C, Contino-Pepin C. Targeting vascular changes in lesions in multiple sclerosis and experimental autoimmune encephalomyelitis. Cent Nerv Syst Agents Med Chem 2012; 12: 7–14. Kobayashi T, Shinoda H, Goto I, Yamanaka T, Suzuki Y. Globoid cell leukodystrophy is a generalized galactosylsphingosine (psychosine) storage disease. Biochem Biophys Res Commun 1987; 144: 41–6. Kobayashi T, Yamanaka T, Jacobs JM, Teixeira F, Suzuki K. The Twitcher mouse: an enzymatically authentic model of human globoid cell leukodystrophy (Krabbe disease). Brain Res 1980; 202: 479–83. Kolter T, Sandhoff K. Sphingolipid metabolism diseases. Biochim Biophys Acta 2006; 1758: 2057–79. Kondo A, Suzuki K. The blood brain barrier in human leukodystrophies and allied diseases. Ultrastructural and morphometric studies on the capillaries in brain biopsies. Clin Neuropathol 1993; 12: 169–74. Kozutsumi Y, Kanazawa T, Sun Y, Yamaji T, Yamamoto H, Takematsu H. Sphingolipids involved in the induction of multinuclear cell formation. Biochim Biophys Acta 2002; 1582: 138–43. Lee HS, Han J, Bai HJ, Kim KW. Brain angiogenesis in developmental and pathological processes: regulation, molecular and cellular communication at the neurovascular interface. FEBS J 2009; 276: 4622–35. Lee JF, Zeng Q, Ozaki H, Wang L, Hand AR, Hla T, et al. Dual roles of tight junction-associated protein, zonula occludens-1, in sphingosine 1-phosphate-mediated endothelial chemotaxis and barrier integrity. J Biol Chem 2006; 281: 29190–200. Li YT, Li SC, Buck WR, Haskins ME, Wu SW, Khoo KH, et al. Selective extraction and effective separation of galactosylsphingosine (psychosine) and glucosylsphingosine from other glycosphingolipids in pathological tissue samples. Neurochem Res 2011; 36: 1612–22. Loonen MC, Van Diggelen OP, Janse HC, Kleijer WJ, Arts WF. Lateonset globoid cell leucodystrophy (Krabbe’s disease). Clinical and genetic delineation of two forms and their relation to the earlyinfantile form. Neuropediatrics 1985; 16: 137–42. MacMillan CJ, Furlong SJ, Doucette CD, Chen PL, Hoskin DW, Easton AS. Bevacizumab diminishes experimental autoimmune encephalomyelitis by inhibiting spinal cord angiogenesis and reducing peripheral T-cell responses. J Neuropathol Exp Neurol 2012; 71: 983–99. Marchesini S, Preti A, Aleo MF, Casella A, Dagan A, Gatt S. Synthesis, spectral properties and enzymatic hydrolysis of fluorescent derivatives of cerebroside sulfate containing long-wavelength-emission probes. Chem Phys Lipids. 1990; 53: 165–75. Ment LR, Stewart WB, Fronc R, Seashore C, Mahooti S, Scaramuzzino D, et al. Vascular endothelial growth factor mediates reactive angiogenesis in the postnatal developing brain. Brain Res Dev Brain Res 1997; 100: 52–61. Mito T, Konomi H, Houdou S, Takashima S. Immunohistochemical study of the vasculature in the developing brain. Pediatr Neurol 1991; 7: 18–22. Nico B, Frigeri A, Nicchia GP, Corsi P, Ribatti D, Quondamatteo F, et al. Severe alterations of endothelial and glial cells in the blood-brain barrier of dystrophic mdx mice. Glia 2003; 42: 235–51.

GALC deficiency and angiogenesis Nozawa M, Iwamoto T, Tokoro T, Eto Y. Novel procedure for measuring psychosine derivatives by an HPLC method. J Neurochem 1992; 59: 607–9. Ogunshola OO, Stewart WB, Mihalcik V, Solli T, Madri JA, Ment LR. Neuronal VEGF expression correlates with angiogenesis in postnatal developing rat brain. Brain Res Dev Brain Res 2000; 119: 139–53. Olsson AK, Dimberg A, Kreuger J, Claesson-Welsh L. VEGF receptor signalling—in control of vascular function. Nat Rev Mol Cell Biol 2006; 7: 359–71. Ribatti D, Gualandris A, Bastaki M, Vacca A, Iurlaro M, Roncali L, et al. New model for the study of angiogenesis and antiangiogenesis in the chick embryo chorioallantoic membrane: the gelatin sponge/ chorioallantoic membrane assay. J Vasc Res 1997; 34: 455–63. Ribatti D, Nico B, Vacca A, Presta M. The gelatin sponge-chorioallantoic membrane assay. Nat Protoc 2006; 1: 85–91. Ronca R, Benzoni P, Leali D, Urbinati C, Belleri M, Corsini M, et al. Antiangiogenic activity of a neutralizing human single-chain antibody fragment against fibroblast growth factor receptor 1. Mol Cancer Ther 2010; 9: 3244–53. Rybak JN, Ettorre A, Kaissling B, Giavazzi R, Neri D, Elia G. In vivo protein biotinylation for identification of organ-specific antigens accessible from the vasculature. Nat Methods 2005; 2: 291–8. Sakai N. Pathogenesis of leukodystrophy for Krabbe disease: molecular mechanism and clinical treatment. Brain Dev 2009; 31: 485–7. Sakai N, Inui K, Tatsumi N, Fukushima H, Nishigaki T, Taniike M, et al. Molecular cloning and expression of cDNA for murine galactocerebrosidase and mutation analysis of the twitcher mouse, a model of Krabbe’s disease. J Neurochem 1996; 66: 1118–24. Segura I, De Smet F, Hohensinner PJ, Ruiz de Almodovar C, Carmeliet P. The neurovascular link in health and disease: an update. Trends Mol Med 2009; 15: 439–51. Smith B, Galbiati F, Castelvetri LC, Givogri MI, Lopez-Rosas A, Bongarzone ER. Peripheral neuropathy in the Twitcher mouse involves the activation of axonal caspase 3. ASN Neuro 2011; 3: pii: e00066. Suzuki K. The twitcher mouse: a model for Krabbe disease and for experimental therapies. Brain Pathol 1995; 5: 249–58. Suzuki K. Twenty five years of the “psychosine hypothesis”: a personal perspective of its history and present status. Neurochem Res 1998; 23: 251–9. Suzuki K. Globoid cell leukodystrophy (Krabbe’s disease): update. J Child Neurol 2003; 18: 595–603. Suzuki K, Suzuki Y. Globoid cell leucodystrophy (Krabbe’s disease): deficiency of galactocerebroside beta-galactosidase. Proc Natl Acad Sci USA 1970; 66: 302–9.

Brain 2013: 136; 2859–2875

| 2875

Svennerholm L, Vanier MT, Mansson JE. Krabbe disease: a galactosylsphingosine (psychosine) lipidosis. J Lipid Res 1980; 21: 53–64. Takahashi H, Igisu H, Suzuki K. Murine globoid cell leukodystrophy: the twitcher mouse. An ultrastructural study of the kidney. Lab Invest 1984; 50: 42–50. Tominaga K, Matsuda J, Kido M, Naito E, Yokota I, Toida K, et al. Genetic background markedly influences vulnerability of the hippocampal neuronal organization in the “twitcher” mouse model of globoid cell leukodystrophy. J Neurosci Res 2004; 77: 507–16. Ueda M, Graf R, MacWilliams HK, Schliwa M, Euteneuer U. Centrosome positioning and directionality of cell movements. Proc Natl Acad Sci USA 1997; 94: 9674–8. Visigalli I, Ungari S, Martino S, Park H, Cesani M, Gentner B, et al. The galactocerebrosidase enzyme contributes to the maintenance of a functional hematopoietic stem cell niche. Blood 2010; 116: 1857–66. Wenger DA, Rafi MA, Luzi P, Datto J, Costantino-Ceccarini E. Krabbe disease: genetic aspects and progress toward therapy. Mol Genet Metab 2000; 70: 1–9. White AB, Galbiati F, Givogri MI, Lopez Rosas A, Qiu X, van Breemen R, et al. Persistence of psychosine in brain lipid rafts is a limiting factor in the therapeutic recovery of a mouse model for Krabbe disease. J Neurosci Res 2011; 89: 352–64. White AB, Givogri MI, Lopez-Rosas A, Cao H, van Breemen R, Thinakaran G, et al. Psychosine accumulates in membrane microdomains in the brain of krabbe patients, disrupting the raft architecture. J Neurosci 2009; 29: 6068–77. Whitfield PD, Sharp PC, Taylor R, Meikle P. Quantification of galactosylsphingosine in the twitcher mouse using electrospray ionizationtandem mass spectrometry. J Lipid Res 2001; 42: 2092–5. Zacchigna S, Lambrechts D, Carmeliet P. Neurovascular signalling defects in neurodegeneration. Nat Rev Neurosci 2008; 9: 169–81. Zaka M, Wenger DA. Psychosine-induced apoptosis in a mouse oligodendrocyte progenitor cell line is mediated by caspase activation. Neurosci Lett 2004; 358: 205–9. Zhu H, Lopez-Rosas A, Qiu X, Van Breemen RB, Bongarzone ER. Detection of the neurotoxin psychosine in samples of peripheral blood: application in diagnostics and follow-up of Krabbe disease. Arch Pathol Lab Med 2012; 136: 709–10. Zlokovic BV. Neurovascular mechanisms of Alzheimer’s neurodegeneration. Trends Neurosci 2005; 28: 202–8.