Gene Therapy (2007) 14, 405–414 & 2007 Nature Publishing Group All rights reserved 0969-7128/07 $30.00 www.nature.com/gt
ORIGINAL ARTICLE
Partial cure of established disease in an animal model of metachromatic leukodystrophy after intracerebral adeno-associated virus-mediated gene transfer C Sevin1, L Verot2, A Benraiss1, D Van Dam3, D Bonnin1, G Nagels3,4, F Fouquet1, V Gieselmann5, MT Vanier2, PP De Deyn3,6, P Aubourg1 and N Cartier1 Institut National de la Sante´ et de la Recherche Me´dicale Inserm U745, Laboratory of Molecular Genetics and Universite´ ParisV, Paris, France; 2Inserm U499, Lyon and Fondation Gillet-Me´rieux, University of Lyon and Lyon-Sud Hospital 69310 Pierre-Benite, France; 3 Laboratory of Neurochemistry and Behavior at Institute Born-Bunge, Department of Biomedical Sciences, University of Antwerp, Antwerp, Belgium; 4Melsbroek and Department of Neurology, National MS Centre, University Hospital, Edegem, Belgium; 5Department of Physiological Chemistry, University of Bonn, Bonn, Germany and 6Department of Neurology and Memory Clinic, Middelheim General Hospital, Antwerp, Belgium 1
Metachromatic leukodystrophy (MLD) is a lysosomal storage disease caused by genetic deficiency of arylsulfatase A (ARSA) enzyme. Failure in catalyzing the degradation of its major substrate, sulfatide (Sulf), in oligodendrocytes and Schwann cells leads to severe demyelination in the peripheral (PNS) and central nervous system (CNS), and early death of MLD patients. The ARSA knockout mice develop a disease that resembles MLD but is milder, without significant demyelination in the PNS and CNS. We showed that adeno-associated virus serotype 5-mediated gene transfer in the brain of ARSA knockout mice reverses Sulf storage and prevents neuropathological abnormalities and
neuromotor disabilities when vector injections are performed at a pre-symptomatic stage of disease. Direct injection of viral particles into the brain of ARSA knockout mice at a symptomatic stage results in sustained expression of ARSA, prevention of Sulf storage and neuropathological abnormalities. Despite these significant corrections, the treated mice continue to develop neuromotor disability. We show that more subtle biochemical abnormalities involving gangliosides and galactocerebroside are in fact not corrected. Gene Therapy (2007) 14, 405–414. doi:10.1038/sj.gt.3302883; published online 9 November 2006
Keywords: lysosomal storage diseases; metachromatic leukodystrophy; adeno-associated virus
Introduction Metachromatic leukodystrophy (MLD) is a lipid storage disorder caused by a deficiency of the lysosomal arylsulfatase A (ARSA; EC 3.1.6.8). In the most common form of the disease (the late infantile form), neurological symptoms start around 18 months of age, progress rapidly, and cause death within 3–5 years.1 Patients with the juvenile and adult forms have a later onset of neurological signs, which initially progress less rapidly, but ultimately lead to the same fatal outcome. ARSA-deficient cells accumulate sphingolipid cerebroside sulfate (sulfatide (Sulf)), which is a major component of myelin, and is involved in myelin formation and cell-to-cell interactions.2,3 In MLD patients, Sulfs accumulate mainly in the microglia, oligodendrocytes and Schwann cells, resulting in widespread and devastating demyelination in the peripheral (PNS) and central Correspondence: Dr C Sevin, Institut National de la Sante´ et de la Recherche Me´dicale Inserm U745 and Universite´ ParisV, Laboratory of Molecular Genetics, 82 avenue Denfert Rochereau, Paris 75014, France. E-mail:
[email protected] Received 10 August 2006; revised 23 September 2006; accepted 25 September 2006; published online 9 November 2006
nervous systems (CNS). However, Sulfs also accumulate in neurons.4 ARSA knockout mice develop a disease that resembles MLD, however, symptoms are less severe.5–7 The clinical signs, which include gait disturbance, impaired motor coordination and hyperactivity, become apparent at around 1 year of age, but they do not shorten the lifespan of the mice. The fact that a milder phenotype is observed in ARSA knockout mice is probably due to the lack of significant cerebral demyelination. One reason for this discrepancy between human patients and mice could be that Sulfs accumulate to a much smaller extent in the brain of ARSA knockout mice than in the brain of MLD patients.8 However, it is also possible that environmental and genetic factors are implicated in the onset and severity of the human disease. Nevertheless, the ARSA knockout mouse is an appropriate model for the investigation of therapeutic interventions. During the past few years, the ARSA knockout mouse has been used to evaluate the efficacy of hematopoietic stem cell gene therapy,9–11 enzyme replacement therapy,12 embryonic stem cell therapy13 and direct transfer of the ARSA gene into the brain.14,15 Each approach has its own specific advantages and pitfalls for use in human therapeutic applications. However, intracerebral delivery of the ARSA gene might be effective more quickly in
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
arresting the rapid demyelinating process that occurs in patients with the infantile form of MLD. We have previously demonstrated that adeno-associated virus serotype 5 (AAV5)-mediated ARSA gene transfer into the brain of ARSA knockout mouse has the capacity to treat neuropathology throughout the entire brain, and prevents motor disability at an early stage of the disease (3 months of age).15 At this age, ARSA knockout mice already present with Sulf accumulation, but still display only minimal clinical and histological symptoms. To be of broad clinical relevance, any new therapeutic intervention would have to be effective against the established disease. Here, we report complete normalization of the neuropathology when intracerebral injections of AAV5-ARSA vector were administered at 6 months of age, when the effects of the disease were already well established. However, intracerebral delivery of the ARSA enzyme had no effect on neuromotor disability, and we show that more subtle biochemical abnormalities were in fact not corrected.
Results Mice treated at the symptomatic stage display widespread, robust and sustained expression of the ARSA enzyme The AAV5-ARSA vector (1.5 109 viral particles/ml, 2 ml/ site of injection) was injected into the cerebellum and internal capsule of 6-month-old ARSA knockout mice (late-treated (LT) group; n ¼ 12 mice). The treated animals were assessed at 9 and 18 months of age, and compared to age-matched: (a) wild-type mice, (b) untreated ARSA knockout mice and (c) ARSA knockout mice in which AAV5-ARSA vector injections had been administered before the onset of symptoms (3 months of age, early-treated (ET) group, see Sevin et al.15). As expected, the highest levels of ARSA expression, assessed by enzyme-linked immunosorbent assay (ELISA), were found in the brain and cerebellum, close to the injection sites (Figure 1). Three months after the injection, mean ARSA levels in the forebrain were similar in the ET and LT mice (404.97117.9 vs 536.97116 ng ARSA/mg protein, P ¼ 0.66; Figure 1a). In both groups, a significant time-dependent increase in ARSA expression was observed at 18 months in the forebrain (ET: 13387130; LT: 17007248 ng ARSA/mg protein; Figure 1a). This increase was associated with caudo-rostral diffusion of the ARSA enzyme from the injection sites to distant areas (Figure 1b). In contrast, levels of ARSA expression in the cerebellum and brainstem did not change over time (Figure 1). A significant but lower level of ARSA was also found in the cervical spinal cord 3 months after injection (ET: 32.6714.6; LT: 13.677.6 ng ARSA/mg protein, P ¼ 0.16) and at 18 months of age (ET: 14.377.2; LT: 8.070.9 ng ARSA/mg protein, P ¼ 0.54; Figure 1a). No recombinant enzyme was detected in the peripheral tissues (liver, spleen, kidney, testis) of treated mice (data not shown). No difference was observed between the histochemical expression patterns of the ARSA enzyme in ET and LT ARSA knockout mice (Table 1). ARSA-positive cells were most abundant in the caudate-putamen, thalamus, superior colliculus and cortex. However, numerous ARSA-positive cells were also detected in the white Gene Therapy
a
2400 ET M6 (n = 3) LT M9 (n = 3) ET M18 (n = 3) LT M18 (n = 3)
NS
ng ARSA/mg protein
406
1800
1200 NS 600
0 Forebrain
Cerebellum Brainstem Spinal cord
b
ng ARSA/mg protein
ET M6
LT M9
ET M18
LT M18
150-200 200-500 500-1000 1000-1500 > 1500
Figure 1 ARSA expression in the whole brain (a) and serial brain sections (b) in ET and LT ARSA knockout mice. ET mice were injected at 3 months and studied at 6 (ET M6) and 18 months (ET M18) of age. LT mice were injected at 6 months and studied at 9 (LT M9) and 18 months (LT M18) of age. (a) Mean level of ARSA expression in the forebrain, cerebellum, brainstem and spinal cord, assessed by ELISA. Numbers in brackets indicate the number of mice for each time point of the study. Black line: baseline ARSA level in human brain, NS: not significant. (b) Time-dependent expression of ARSA. The gray grading indicates the level of expression of ARSA. White stars indicate the injection sites.
Table 1 Quantification of ARSA-positive cells in different areas of the brain in ET and LT ARSA knockout mice, 3 months after injection and at the end of the study (18 months of age) Localization
Olfactory bulbs Cortex Superior colliculus Inferior colliculus Corpus callosum Internal capsule Hippocampus Thalamus Caudate Hypothalamus Cerebellar white matter Cerebellar cortex Pons
Three months post-injection End of the study ET
LT
ET
LT
+/ +++ +++ + + +/ ++ ++++ +++ ++ ++ ++++ ++
+ +++ ++ + + +/ ++ +++ ++++ + ++ +++ ++
+ +++ ++ ++ ++ + ++ ++++ +++ +++ ++ ++++ +++
+ +++ ++ ++ ++ +/ ++ +++ +++ ++ ++ ++++ +++
Abbreviations: ARSA, arylsulfatase A; ET, early treated; LT, late treated. The grading indicates the number of ARSA-positive cells in sections of early- and late-treated ARSA knockout mice. (++++: 4200 cells per microscope section at 10 magnification; +++: 100–200 cells; ++: 30–100 cells; +: 10–30 cells; +/: 1–10 cells); n ¼ 10 sections per mouse (three mice/group for each point of the study) were examined for each site.
matter of the corpus callosum and fimbria. These cells displayed the morphology of microglial cells or astrocytes. In the cerebellum, ARSA staining was observed in the cortex and in the white matter. A time-dependent increase in the number of ARSA-positive cells was observed in the corpus callosum, inferior colliculus,
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
hypothalamus and pons, suggesting that human ARSA enzyme was being secreted by transduced cells and recaptured by remote non-transduced cells.
Double immunostaining showed that most ARSAexpressing cells were neurons (Figure 2a), but the ARSA enzyme was also expressed in astrocytes (Figure 2b) and, to a lesser extent, in microglial cells (Figures 2c). ARSApositive cells showed punctate, perinuclear staining indicating that the recombinant enzyme was located in the lysosomes (Figure 2, arrows). No ARSA expression could be detected in the oligodendrocytes.
407
ARSA delivery prevents Sulf storage and improves the Sulf/GalC ratio Sulf storage is already detectable in 3-month-old ARSA knockout mice using Alcian blue staining and it increases thereafter with age.4,15 In all treated ARSA knockout mice (ET and LT), Sulf storage had been abolished or at least markedly reduced, within 3 months after injection and remained unchanged thereafter, which contrasted with the age-dependent increase of Sulf storage seen in the untreated mice (Figure 3). Faint residual staining was sometimes observed in macrophages localized in the white matter of cerebellum (Figure 3, arrows). Sulf storage was apparently absent in cells of treated mice showing the morphology of oligodendrocytes. To determine the Sulf content in treated mice more precisely, we measured the Sulf/galactosylceramide (GalC) ratio, variations of which are a sensitive indicator of ARSA activity (Figure 4). In the wild-type mice, the Sulf/GalC ratio showed a slight continuous increase between 3 and 18 months of age (+20% in the brain; +40% in the cerebellum, brainstem and spinal cord). In the untreated ARSA knockout mice, this ratio, already significantly higher than normal at 3 months of age, showed a steep and almost linear increase (+100% in the brain, cerebellum and brainstem; +80% in the spinal cord) during the same period. By contrast, in the ET ARSA knockout mice, the Sulf/GalC ratio remained essentially stable throughout the entire study period in the brain, brainstem and cerebellum, with a pattern following the same kinetics as observed in the wild-type mice. The same phenomenon was observed in the brain and brainstem of the LT mice. However, the Sulf/GalC ratios were not completely corrected in the cerebellum of the LT mice, or in the spinal cord of either ET or LT animals at 18 months of age. ARSA delivery decreases the Sulf storage, but does not increase GalC The increase in the Sulf/GalC ratio observed in both MLD patients and mice could reflect an increase in Sulf but also a decrease in GalC contents, the latter being an indicator of myelin/oligodendrocyte integrity.16,17 To better evaluate the participation of Sulf and GalC in the observed changes, we measured the GalC level in the forebrain of 18-month-old wild-type, untreated and treated ARSA knockout mice (n ¼ 2 mice/group). The
Figure 2 Double-immunostaining with antibody against human ARSA and cellular markers for neurons (NeuN, a), astrocytes (GFAP, b) and microglia (tomato-lectin, c) in the cerebral cortex of LT ARSA knockout mice. Recombinant enzyme shows a punctate perinuclear staining suggesting a lysosomal localization (white arrows). Gene Therapy
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
408
Figure 3 Sulf storage (assessed using Alcian blue staining) in the fimbria and cerebellum of wild-type (WT), untreated (UT), and ET and LT ARSA knockout mice. Arrows: residual staining observed in the cerebellar white matter of LT mice.
concentration of GalC was significantly decreased in the untreated ARSA knockout mice (11.1 vs 17.3 mmol GalC/ g wet weight in controls). In the ET and LT ARSA knockout mice, GalC concentrations remained essentially similar to the values observed in untreated animals (12.4 and 9.7 vs 11.1 mmol GalC/g wet weight).
The brain delivery of ARSA enzyme does not correct the abnormal accumulation of gangliosides A variable level of accumulation of some gangliosides has been shown to occur in several lysosomal storage diseases (LSDs), not primarily involving the degradation of gangliosides. The ganglioside patterns (GM3, GM2, GM1, GD3, GD2, GD1a, GD1b; GT1; GQ1) were quantified in the brains of wild-type mice as well as treated and untreated ARSA knockout animals at 12 and 18 months of age (Figure 5). A twofold increase in the proportion of GM2 ganglioside was observed in the 12-month-old ARSA knockout mice (Figure 5a), whereas the level of GD3 was increased Gene Therapy
by 30% (Figure 5b). GM2 accumulation remained stable between 12 and 18 months of age, whereas that of GD3 continued to increase. No significant modification of any other gangliosides was detected (data not shown). In the 18-month-old ET and LT mice, GM2 ganglioside accumulation appeared partially corrected (Figure 5a), whereas the proportion of GD3 remained unchanged (Figure 5b).
Both ET and LT mice are protected against brain damage The 18-month-old ARSA knockout mice display glycolipid storage (Periodic Acid Schiff (PAS)-reactive material), microglial activation (increased number of swollen ameboid cells) and severe astrogliosis, mainly in the white matter (see Sevin et al.15). These abnormalities were fully prevented in both the ET and LT ARSA knockout mice (data not shown). The neuronal degeneration of the Purkinje cells observed in 18-month-old untreated animals was also
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
409 WT UT ET LT
forebrain
cerebellum 0.8
Sulf/GalC ratio
Sulf/GalC ratio
0.8 0.6 0.4 0.2 0
0.6 0.4 0.2 0
0
3
6
9
12
15
0
18
3
6
months brainstem
12
15
18
15
18
spinal cord 0.8
Sulf/GalC ratio
0.8
Sulf/GalC ratio
9
months
0.6 0.4 0.2 0
0.6 0.4 0.2 0
0
3
6
9
12
15
18
months
0
3
6
9
12
months
Figure 4 Sulf/GalC ratio in the forebrain, cerebellum, brainstem, and spinal cord of wild-type (WT), untreated (UT), ET and LT ARSA knockout mice. Three mice were studied at each time point of the study.
markedly improved in mice treated at an early or late stage of the disease (Figure 6a and b).
LT ARSA knockout mice are not protected against neuromotor impairment Results of rotarod performances of ET mice have been published, demonstrating that intracerebral injections of AAV5-ARSA vector markedly improve neuromotor performances (see Sevin et al.15 and Figure 7). However, when the AAV vector was injected at a symptomatic stage of the disease, it induced no amelioration of the neuromotor disabilities observed in ARSA knockout mice (Figure 7).
Discussion The use of viral vectors for gene delivery into the nervous system shows great promise for therapeutic applications, and the preclinical development of gene transfer for the treatment of neurological disorders is proceeding rapidly. As lysosomal enzymes can be secreted by transduced cells and recaptured by nontransduced cells via the mannose-6-phosphate (M6P) receptor pathway, gene therapy for LSDs could theoretically provide a specific and long-term solution (reviewed by Sands and Davidson18). MLD is one of these LSDs for which there is currently no effective treatment; the benefits of allogeneic hematopoietic stem cell transplantation are still controversial.19 Even though ARSA enzyme deficiency is systemic, manifestations of the disease are restricted to the nervous system. Thus, the rationale behind a CNS gene therapy approach in
MLD is straightforward. However, in contrast to many other LSDs, the enzyme deficiency in MLD affects not only neurons but also mainly glial cells in the white matter (in particular oligodendrocytes), and Schwann cells in the peripheral nerves. In fact, the human disease is characterized primarily by myelin degeneration in the CNS and PNS. Although enzyme replacement therapy is a promising therapeutic option for the treatment of peripheral nerve in MLD,12 CNS gene therapy is much more challenging. Owing to the inability of current gene therapy vectors to transduce oligodendrocytes in vivo,20 the bioavailabity of functional ARSA enzyme in oligodendrocytes is more difficult to achieve, depending among other things on the abundance of M6P receptors, and the level of endocytosis activity in these cells. In a previous study, we have shown that AAV5mediated ARSA gene tranfer can be used to treat the CNS damage in ARSA knockout mice, and prevent the neuromotor disability when ARSA gene delivery is performed at an early stage of the disease, that is, at 3 months of age in this animal model of MLD.15 In particular, we demonstrated the widespread diffusion of ARSA, and an increase in the number of cells containing the recombinant enzyme over time, indicating that the ARSA enzyme was probably secreted by transduced cells and recaptured by non-transduced cells. It was noteworthy, however, that expression of the ARSA enzyme was mostly restricted to neurons, with only a few astrocytes and microglial cells expressing the ARSA enzyme. In the present study, AAV5-mediated ARSA delivery was performed in symptomatic ARSA knockout mice (at 6 months of age). ARSA gene delivery led to the same Gene Therapy
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
410
a 2.5
WT UT ET LT
GM2
2
a
1.5
1
0.5 12
18 months * 20
* 15 10 5
WT
0
3
12
18 months
Figure 5 Proportions of GM2 (a) and GD3 (b) gangliosides in the forebrain of wild-type (WT), untreated (UT), and ET and LT ARSA knockout mice. Results are expressed as molar percentage of total gangliosides. Three mice were studied at each time point of the study.
normalization of the neurological damage that included correction of astrogliosis and microglial activation, and the rescue of Purkinje cells. The intracellular Sulf detoxification may have interrupted a degenerative cascade, and thus reconstituted normal functions in Purkinje cells. There was also evidence of time-dependent increase in the level of ARSA expression in the brain, indicating that the ARSA enzyme was secreted and recaptured. As we did not investigate the expression of M6P receptors in CNS cells, or M6P residues on recombinant ARSA, we cannot exclude the possibility that mechanisms other than M6P-mediated uptake could have mediated the enzyme transfer. The ARSA enzyme, like other lysosomal enzymes, undergoes axonal transport both in vitro21 and in vivo.22 The time required for maximal diffusion of ARSA to occur could be explained by the large distance – several millimeters – that the ARSA enzyme has to cross to reach areas located far from the injection sites. Another explanation could be impaired motility of the lysosomes along axons due to an increase in ganglioside content (see below). The absence of any improvement of neuromotor disability in the ARSA knockout mice treated at a symptomatic stage is puzzling. Similar observations have been reported in animal models of Krabbe disease, Gene Therapy
Figure 6 Prevention of Purkinje cell degeneration. (a) Eighteenmonth-old ARSA knockout mice display Purkinje cell loss (white arrows) with simplified dendritic architecture, which are mostly prevented in ET and LT ARSA knockout mice (black arrows). Scale bar: 100 mm (b) quantification of Purkinje cells in cerebellum of 18-month-old wild-type (WT), untreated (UT), and ET and LT ARSA knockout mice. A significant decrease in the number of Purkinje cells was observed in UT animals that was mostly prevented in ET and LT mice. ***Po0.01 between WT and UT; *P ¼ 0.05 between LT and UT and between ET and UT.
Intracerebral injection of AAV vector ET LT
300
Time on road during 4th trial (s)
2
LT
4
***
ET
GD3
5
25
UT
b
6
Number of Purkinje cells / mm
b
250 200 x
150 100 50
WT UT ET LT
0 3
6
9
12
15
18
months Figure 7 Motor performance in wild-type (WT), untreated (UT), and ET and LT ARSA knockout mice. Data on mice treated at a pre-symptomatic stage (ET) have been published by Sevin et al.15 Arrows indicate the time of the treatment in ET and LT animals.
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
a leukodystrophy caused by mutations in the galactocerebrosidase gene.23 The galactocerebrosidase enzyme (GALC) degrades various galactolipids in the lysosomes, and is important for the normal myelination of axons. Recombinant AAV1, 2 and 5 have been used to transfer the GALC gene into the brain of Krabbe mouse models, which resulted in sustained expression of the GALC enzyme. Although the treatment did improve myelination, it only slightly prolonged the life expectancy of the treated mice.20,24 Neurophysiological PNS dysfunction has been demonstrated in ARSA knockout mice with a C57BL6/129 mixed genetic background.9,12 We performed studies of nerve conduction velocities, F-wave latency and somatosensory-evoked potentials in wild-type, treated and untreated ARSA knockout mice at 18 months of age but failed to detect any abnormalities (data not shown). As we did not investigate motor-evoked potentials, we cannot exclude the possibility that the spinal pyramidal tracts could be involved in ARSA knockout mice. However, this is very unlikely, because the sensory and motor pathways in the spinal cord are usually affected to the same extent by demyelination in all leukodystrophies. We hypothesize that the lack of beneficial effects on neuromotor disability in LT ARSA knockout mice is more likely due to subtle abnormalities in the brain that were not corrected by ARSA gene transfer. Alcian blue staining showed that Sulf storage was abolished or markedly reduced. Biochemical studies confirmed that the Sulf accumulation was largely prevented in the brain, although to a lesser extent in the cerebellum and upper cervical spinal cord. GalC, a major and relatively specific lipid component of myelin, plays critical roles in the development and maintenance of myelin.2,16,17 We observed a 36% decrease in the GalC content in the forebrain of 18-month-old untreated ARSA knockout mice, in line with data reported by others.4 However, in the AAV5-ARSAtreated mice, the GalC levels did not increase, the mice treated at a symptomatic stage having the lowest levels. The failure to restore normal GalC levels might reflect a mild degree of demyelination, which is not detectable histologically, but is sufficient to impair motor functions. A secondary accumulation of the minor monosialogangliosides GM2 and GM3 occurs in a number of LSDs, including Niemann–Pick A/B and C, and mucopolysaccharidosis I and III diseases.25 Several gangliosides, among which GM2, GD3, GM3 and GD2 have been reported elevated in the white matter of patients suffering from various leukodystrophies, including MLD and Krabbe disease.26,27 Similarly, ARSA knockout mice accumulate GM2 and GD3 gangliosides. An abnormal accumulation of GM2 ganglioside has been implicated in impaired dendritogenesis and synaptogenesis, abnormal neurite growth and neuronal apoptosis.28–30 GD3 is usually expressed at high levels in activated microglia and reactive astrocytes.25 It is possible that the accumulation of gangliosides and other unidentified metabolites could also impair the trafficking of lysosomes from neuronal somata to the axonal and dendritic synapses.21,31 A decrease in GM2 content was observed in ARSA knockout mice treated at an early stage (30% decrease) or at the symptomatic stage (15%), but this was probably not sufficient to reach a therapeutic effect. The GD3
increase was not corrected, even though astrogliosis and microglial activation were not detected by neuropathological studies of the treated mice. ARSA enzyme requires post-translational modifications and oxidation of a conserved cysteine residue by formylglycine-generating enzyme (FGE) to be active.32,33 Coexpression of FGE and ARSA enzyme might enhance the therapeutic effect of ARSA gene delivery.34 As we used an AAV vector encoding the human ARSA cDNA to treat the ARSA knockout mice, subtle differences in posttranslational modifications between the human and the mouse ARSA enzyme and non-optimal activation of ARSA by FGE enzyme could also have contributed to the lack of complete correction in ARSA knockout mice treated at a symptomatic stage. Although additional initiatives may be required to reverse the course of the disease at a symptomatic stage, the results of intracerebral ARSA gene transfer on established disease of ARSA knockout mice remain encouraging for the future application of CNS gene therapy in MLD. This reminds us that the ‘proof of concept’ of any new treatment, particularly gene therapy, must not be limited to a demonstration that a disease or even clinical abnormalities can be corrected, particularly in mouse models. In fact, for most neurodegenerative CNS diseases, we are far from having a full comprehensive understanding of the disorder. We need to find more sensitive markers of disease pathology that may prove useful not only for assessing the efficacy of new treatments using animal models but also for identifying the clinical benefits that can be expected from phaseI/II trials in patients.
411
Materials and methods Production of the AAV vector The pAAV5-ARSA plasmid encoding the human ARSA cDNA was generated as described previously.15 AAV5ARSA viral stocks were generated by the transient transfection of 293 cells,35 and purified using CsCl ultracentrifugation gradients. The final titer was 1.5 1012 physical particles/ml. Animal models ARSA knockout mice with a mixed 129sv/Ola genetic background were bred from homozygous founders.5 Animal experiments were approved by the veterinary bureau of the Institut National de la Sante´ et de la Recherche Me´dicale (Inserm) and the Animal Ethics Committee of the University of Antwerp, and were carried out in compliance with the Guide for the Care and Use of Laboratory animals (NIH publication no. 85-24, revised) and the European Communities Council Directive 86/609/EEC. Direct injection of the AAV vector into the brain of ARSA knockout mice ARSA knockout mice were anesthetized by intraperitoneal injection of ketamine/xylazine (0.1/0.05 mg/g body weight), and positioned on a stereotactic frame (David Kopf Instruments, Tujunga, CA, USA). Two microliters of AAV5-ARSA vector preparation (3 109 physical particles of AAV5-ARSA) were injected into the cerebellar vermis (stereotactic coordinates from the bregma: AP: Gene Therapy
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
412
6.48; ML: 0; DV: 1.6 mm, according to Franklin and Praxinos36), and into the left and the right internal capsules (AP: 1.2 mm; ML: 72.0 mm; DV: 3.8 mm), using a 30-gauge blunt micropipette attached to a 10 ml Hamilton syringe (Reno, NV, USA).
Behavior testing The animals were housed in standard mouse cages under conventional laboratory conditions with food and water available ad libitum and a 12/12 h light–dark cycle. Treated (n ¼ 9–10) and untreated ARSA knockout mice (n ¼ 16–25) and wild-type mice (n ¼ 10–12) were tested at 3, 6, 9, 12, 15 and 18 months. Balance and motor coordination were evaluated on an accelerating rotarod apparatus (Ugo Basile, Italy). After two adaptation trials lasting up to 2 min each at a constant speed (4 r.p.m.), the mice were placed on the rotating rod for four test trials during which the rotation speed was gradually increased from 4 to 40 r.p.m. The time an animal was able to stay on the rod was measured up to a maximum of 5 min. Differences in rotarod performance were assessed using a two-way analysis of variance (ANOVA) with correction for repeated measures (2-way RM-ANOVA). Tukey post hoc multiple comparisons were used to analyze age-dependent differences between treatment groups for specific parameters. Statistical analysis was performed using Sigmastat software (SPSS Inc., Erkrath, Germany), with the level of probability set at 95%. Tissue processing Anesthetized animals were transcardially infused with phosphate-buffered saline (PBS), in some cases followed by 4% paraformaldehyde (PFA) in PBS for histological studies. The brain, spinal cord, liver, spleen and testis were removed. For molecular and biochemical analysis (n ¼ 3 mice for each point of the study), brain samples were placed on a mouse coronal brain matrix (World Precision Instrument Ltd, Stevenage, UK) and 1-mm thick coronal brain sections were prepared. Samples were immediately frozen and stored at 801C. For histological and immunohistological studies (n ¼ 3 mice for each point of the study), the tissues were post-fixed in 4% PFA, cryoprotected in 30% sucrose, embedded in OCT (Tissutek, Durham, NC) at 551C, and cut into 10-mm sections. Determination of ARSA expression Tissues were homogenized in lysis buffer, sonicated and centrifuged. The supernatants were collected for determination of the protein content, and the concentration of recombinant ARSA was immunologically measured by an indirect sandwich ELISA, as described.15,37 Results were expressed as mean7s.e.m. of three assays. Brain lipids analysis Total lipids were extracted from cerebral slices (7–50 mg) in chloroform–methanol–water 4:8:3 (by volume) as described.38 Each brain slice had been weighed before freezing, to avoid weight error owing to dehydration of the sample. A suitable aliquot of the total lipid extract was used for determination of the Sulf/GalC ratios. Lipids were subjected to alkaline saponification and desalted by phase partition and samples were applied to high-performance silica gel 60 thin-layer chromatograGene Therapy
phy plates (HPTLC) (Merck, Darmstadt, Germany) using a Camag linomat IV apparatus. Plates were developed in chloroform–methanol–water 65:25:4 (by volume). Detection with orcinol-sulfuric acid was performed using the Camag chromatogram immersion device III, and densitometry at 520 nm using a CAMAG TLC scanner model II. Data were expressed as mean7s.e.m. of the Sulf/GalC ratios.15 For determination of the GalC contents, similar aliquots from the total lipid extract derived from three brain hemisphere slices with the same topology were pooled. As the GalC concentration was directly correlated to the proportion of white matter in each slice, this strategy was adopted to best minimize differences owing to different topography of the brain tissue analyzed. Quantitative HPTLC was performed using the chromatography system and devices described above, and a GalC standard prepared from human brain in the laboratory. Results were expressed as mmol/g wet weight. For ganglioside determinations, the same method of lipid extraction38 was used. For the 12-month-old wildtype and untreated ARSA mice, each analysis was conducted on one forebrain hemisphere. For the 18month-old mice, similar aliquots from three forebrain slices with identical topography were pooled. Gangliosides were isolated using Bond–Elut C-18 reverse phase columns (Varian) according to Kyrklund.39 This procedure allows clean separation of acidic lipids from other lipids without loss of the less polar gangliosides and simultaneous desalting. Separation of the individual gangliosides from an aliquot corresponding to 2.5–3 mg fresh tissue was performed on silica gel 60 HPTLC plates (see above) developed in chloroform–methanol–0.2% CaCl2 (55:45:10, v/v/v). Gangliosides were visualized by the resorcinol-HCl reagent and quantified by densitometry at 580 nm using the Camag scanner. Results were expressed as molar percentage of the individual gangliosides, taking into account the number of sialic acids per molecule.
Histopathology Standard histological procedures. Frozen brain sections were stained with PAS according to a standard procedure. Sulf storage was evaluated using Alcian blue staining as described.40 Immunohistochemistry. Immunostaining of ARSA was performed as described.15 Double immunostaining with anti-ARSA and cellular markers was carried out using the following antibodies: anti-NeuN for neurons (Chemicon International, Temecula, CA, USA), anti-glial fibrillary acidic protein for astrocytes (GFAP, Abcam, Cambridge, MA, USA), anti-CNPase for oligodendrocytes (Sigma-Aldrich), Tomato Lectin for microglia (Sigma-Aldrich, Lyon, France). Cell counts. The Purkinje cell count was performed using Histolab image analyzer software (Microvision Instruments, Paris, France). For each mouse, 10 brain sections were analyzed by two independent investigators blind with regard to treatment status of the animals.
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
413
Abbreviations AAV, adeno-associated virus; ARSA, arylsulfatase A; CNS, central nervous system; ELISA, enzyme-linked immunosorbent assay; ET, early treated; GalC, galactosylceramide; GALC, galactocerebrosidase enzyme; GFAP, glial fibrillary acidic protein; LSDs, lysosomal storage diseases; LT, late treated; MLD, metachromatic leukodystrophy; M6P, mannose-6-phosphate; PAS, Periodic Acid Schiff; PFA, paraformaldehyde; PNS, peripheral nervous system; Sulf, sulfatide.
9
10
11
Acknowledgements We thank A Salvetti, P Moullier and the vector core of the Association Franc¸aise contre les Myopathies in Nantes for the preparation of the AAV vector. We express our gratitude to R D’Hooge who contributed to the initial phase of this work. We also thank Frieda Franck (Antwerp, Belgium) for her excellent collaboration and J Lopez for her technical assistance. This work was supported by grants from the European Leukodystrophy Foundation (ELA), the National Organization for Rare Disorders (NORD), the Association Franc¸aise contre les Myopathies (AFM), the GIS-Institut des maladies rares, the Fe´de´ration pour la Recherche sur le Cerveau, the Fondation pour la Recherche Me´dicale (FRM) and the Fund for Scientific Research-Flanders (FWO grant G.0038.05), agreement between the Institute Born-Bunge and the University of Antwerp, Neurosearch Antwerp, Antwerp Medical Research Foundation, the Thomas Riellaerts Fund. DVD holds a postdoctoral position at the University of Antwerp.
14
References
17
1 von Figura K, Gieselmann V, Jaeken J. Metachromatic leukodystrophy. In: Scriver CR, Beaudet AL, Sly WS and Valle D (eds). The Metabolic & Molecular Bases of Inherited Disease, 8th edn, McGraw-Hill Published: New York, 2001, pp 3695–3724. 2 Dupree JL, Popko B. Genetic dissection of myelin galactolipid function. J Neurocytol 1999; 28: 271–279. 3 Pernber Z, Molander-Melin M, Berthold CH, Hansson E, Fredman P. Expression of the myelin and oligodendrocyte progenitor marker sulfatide in neurons and astrocytes of adult rat brain. J Neurosci Res 2002; 69: 86–93. 4 Molander-Melin M, Pernber Z, Franken S, Gieselmann V, Mansson JE, Fredman P. Accumulation of sulfatide in neuronal and glial cells of arylsulfatase A deficient mice. J Neurocytol 2004; 33: 417–427. 5 Hess B, Saftig P, Hartmann D, Coenen R, Lullmann-Rauch R, Goebel HH et al. Phenotype of arylsulfatase A-deficient mice: relationship to human metachromatic leukodystrophy. Proc Natl Acad Sci USA 1996; 93: 14821–14826. 6 D’Hooge R, Coenen R, Gieselmann V, Lullmann-Rauch R, De Deyn PP. Decline in brainstem auditory-evoked potentials coincides with loss of spiral ganglion cells in arylsulfatase A-deficient mice. Brain Res 1999; 847: 352–356. 7 D’Hooge R, Hartmann D, Manil J, Colin F, Gieselmann V, De Deyn PP. Neuromotor alterations and cerebellar deficits in aged arylsulfatase A-deficient transgenic mice. Neurosci Lett 1999; 273: 93–96. 8 Saravanan K, Schaeren-Wiemers N, Klein D, Sandhoff R, Schwarz A, Yaghootfam A et al. Specific downregulation and
12
13
15
16
18 19
20
21
22
23 24
25
mistargeting of the lipid raft-associated protein MAL in a glycolipid storage disorder. Neurobiol Dis 2004; 16: 396–406. Biffi A, De Palma M, Quattrini A, Del Carro U, Amadio S, Visigalli I et al. Correction of metachromatic leukodystrophy in the mouse model by transplantation of genetically modified hematopoietic stem cells. J Clin Invest 2004; 113: 1118–1129. Matzner U, Hartmann D, Lullmann-Rauch R, Coenen R, Rothert F, Mansson JE et al. Bone marrow stem cell-based gene transfer in a mouse model for metachromatic leukodystrophy: effects on visceral and nervous system disease manifestations. Gene Therapy 2002; 9: 53–63. Matzner U, Harzer K, Learish RD, Barranger JA, Gieselmann V. Long-term expression and transfer of arylsulfatase A into brain of arylsulfatase A-deficient mice transplanted with bone marrow expressing the arylsulfatase A cDNA from a retroviral vector. Gene Therapy 2000; 7: 1250–1257. Matzner U, Herbst E, Hedayati KK, Lullmann-Rauch R, Wessig C, Schroder S et al. Enzyme replacement improves nervous system pathology and function in a mouse model for metachromatic leukodystrophy. Hum Mol Genet 2005; 14: 1139–1152. Klein D, Schmandt T, Muth-Kohne E, Perez-Bouza A, Segschneider M, Gieselmann V et al. Embryonic stem cell-based reduction of central nervous system sulfatide storage in an animal model of metachromatic leukodystrophy. Gene Therapy 2006 (Epub ahead of print). Consiglio A, Quattrini A, Martino S, Bensadoun JC, Dolcetta D, Trojani A et al. In vivo gene therapy of metachromatic leukodystrophy by lentiviral vectors: correction of neuropathology and protection against learning impairments in affected mice. Nat Med 2001; 7: 310–316. Sevin C, Benraiss A, Van Dam D, Bonnin D, Nagels G, Verot L et al. Intracerebral adeno-associated virus-mediated gene transfer in rapidly progressive forms of metachromatic leukodystrophy. Hum Mol Genet 2006; 15: 53–64. Jurevics H, Hostettler J, Muse ED, Sammond DW, Matsushima GK, Toews AD et al. Cerebroside synthesis as a measure of the rate of remyelination following cuprizone-induced demyelination in brain. J Neurochem 2001; 77: 1067–1076. Muse ED, Jurevics H, Toews AD, Matsushima GK, Morell P. Parameters related to lipid metabolism as markers of myelination in mouse brain. J Neurochem 2001; 76: 77–86. Sands MS, Davidson BL. Gene therapy for lysosomal storage diseases. Mol Ther 2006; 13: 839–849. Peters C, Steward CG. Hematopoietic cell transplantation for inherited metabolic diseases: an overview of outcomes and practice guidelines. Bone Marrow Transplant 2003; 31: 229–239. Rafi MA, Zhi Rao H, Passini MA, Curtis M, Vanier MT, Zaka M et al. AAV-mediated expression of galactocerebrosidase in brain results in attenuated symptoms and extended life span in murine models of globoid cell leukodystrophy. Mol Ther 2005; 11: 734–744. Chen F, Vitry S, Hocquemiller M, Desmaris N, Ausseil J, Heard JM. alpha-L-Iduronidase transport in neurites. Mol Genet Metab 2006; 87: 349–358. Luca T, Givogri MI, Perani L, Galbiati F, Follenzi A, Naldini L et al. Axons mediate the distribution of arylsulfatase A within the mouse hippocampus upon gene delivery. Mol Ther 2005; 12: 669–679. Suzuki K. Globoid cell leukodystrophy (Krabbe’s disease): update. J Child Neurol 2003; 18: 595–603. Lin D, Fantz CR, Levy B, Rafi MA, Vogler C, Wenger DA et al. AAV2/5 vector expressing galactocerebrosidase ameliorates CNS disease in the murine model of globoid-cell leukodystrophy more efficiently than AAV2. Mol Ther 2005; 12: 422–430. d’Azzo A, Tessitore A, Sano R. Gangliosides as apoptotic signals in ER stress response. Cell Death Differ 2006; 13: 404–414. Gene Therapy
AAV-mediated gene transfer in metachromatic leukodystrophy C Sevin et al
414 26 Svennerholm L, Vanier MT. Brain gangliosides in Krabbe disease. In: Aronson S, Volk BW (eds). Advance in Experimental Medicine and Biology. Plenum Press: New York, 1972, pp 499–514. 27 Suzuki K. Ganglioside patterns of normal and pathological brains. In: Aronson S, Volk BW (eds). Inborn Disorders of Sphingolipid Metabolism. Pergamon Press: Oxford, 1967, pp 215–230. 28 Huang JQ, Trasler JM, Igdoura S, Michaud J, Hanal N, Gravel RA. Apoptotic cell death in mouse models of GM2 gangliosidosis and observations on human Tay-Sachs and Sandhoff diseases. Hum Mol Genet 1997; 6: 1879–1885. 29 Desmaris N, Verot L, Puech JP, Caillaud C, Vanier MT, Heard JM. Prevention of neuropathology in the mouse model of Hurler syndrome. Ann Neurol 2004; 56: 68–76. 30 Buccoliero R, Futerman AH. The roles of ceramide and complex sphingolipids in neuronal cell function. Pharmacol Res 2003; 47: 409–419. 31 Ito JI, Young ZL, Masuda-Isobe M, Tanaka R. Suppression by gangliosides of polymerization of glial cytoskeletons prepared from rat astrocytes: a role of sialic acid moiety. Neurochem Int 1997; 31: 525–531. 32 Dierks T, Dickmanns A, Preusser-Kunze A, Schmidt B, Mariappan M, von Figura K et al. Molecular basis for multiple sulfatase deficiency and mechanism for formylglycine generation of the human formylglycine-generating enzyme. Cell 2005; 121: 541–552. 33 Roeser D, Preusser-Kunze A, Schmidt B, Gasow K, Wittmann JG, Dierks T et al. A general binding mechanism for all human
Gene Therapy
34
35
36 37
38
39
40
sulfatases by the formylglycine-generating enzyme. Proc Natl Acad Sci USA 2006; 103: 81–86. Takakusaki Y, Hisayasu S, Hirai Y, Shimada T. Coexpression of formylglycine-generating enzyme is essential for synthesis and secretion of functional arylsulfatase A in a mouse model of metachromatic leukodystrophy. Hum Gene Ther 2005; 16: 929–936. Salvetti A, Oreve S, Chadeuf G, Favre D, Cherel Y, ChampionArnaud P et al. Factors influencing recombinant adeno-associated virus production. Hum Gene Ther 1998; 9: 695–706. Franklin BJ, Praxinos GT. The Mouse Brain in Stereotaxic Coordinates. Academic Press: New York, 1996. Matzner U, Schestag F, Hartmann D, Lullmann-Rauch R, D’Hooge R, De Deyn PP et al. Bone marrow stem cell gene therapy of arylsulfatase A-deficient mice, using an arylsulfatase A mutant that is hypersecreted from retrovirally transduced donor-type cells. Hum Gene Ther 2001; 12: 1021–1033. Fujita N, Suzuki K, Vanier MT, Popko B, Maeda N, Klein A et al. Targeted disruption of the mouse sphingolipid activator protein gene: a complex phenotype, including severe leukodystrophy and wide-spread storage of multiple sphingolipids. Hum Mol Genet 1996; 5: 711–725. Kyrklund T. Two procedures to remove polar contaminants from a crude brain lipid extract by using prepacked reversed-phase columns. Lipids 1987; 22: 274–277. Schott I, Hartmann D, Gieselmann V, Lullmann-Rauch R. Sulfatide storage in visceral organs of arylsulfatase A-deficient mice. Virchows Arch 2001; 439: 90–96.
