Research Articles: Neurobiology of Disease Fingolimod rescues demyelination in a mouse model of Krabbe's disease https://doi.org/10.1523/JNEUROSCI.2346-19.2020
Cite as: J. Neurosci 2020; 10.1523/JNEUROSCI.2346-19.2020 Received: 30 September 2019 Revised: 17 December 2019 Accepted: 21 January 2020
This Early Release article has been peer-reviewed and accepted, but has not been through the composition and copyediting processes. The final version may differ slightly in style or formatting and will contain links to any extended data.
Alerts: Sign up at www.jneurosci.org/alerts to receive customized email alerts when the fully formatted version of this article is published.
Copyright © 2020 Be´chet et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, which permits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.
1 Fingolimod rescues demyelination in a mouse model of Krabbe’s disease
2 Sibylle Béchet, Sinead O’Sullivan, Justin Yssel, Steven G. Fagan, Kumlesh K. Dev
3 Drug Development, School of Medicine, Trinity College Dublin, IRELAND
4
5 Corresponding author: Prof. Kumlesh K. Dev
6 Corresponding author’s address: Drug Development, School of Medicine, Trinity College
7 Dublin, IRELAND, D02 R590
8 Corresponding author’s phone and fax: Tel: +353 1 896 4180
9 Corresponding author’s e-mail address: [email protected]
10
11 Number of pages: 35 pages
12 Number of figures: 9 figures
13 Number of words (Abstract): 216 words
14 Number of words (Introduction): 641 words
15 Number of words (Discussion): 1475 words
16
17 Abbreviated title: Investigating the use of FTY720 in Krabbe’s disease
18
19 Acknowledgement statement (including conflict of interest and funding sources): This work
20 was supported, in part, by Trinity College Dublin, Ireland and the Health Research Board,
21 Ireland. S.B. is a Ph.D scholar of Trinity College Dublin. We would like to thank Dr. Catherine
22 O’Sullivan (Trinity College Dublin) for aiding in the earlier parts of this study. The authors
23 declare no competing financial interests.
24 Abstract
25 Krabbe’s disease is an infantile neurodegenerative disease, affected by mutations in the
26 lysosomal enzyme galactocerebrosidase, leading to the accumulation of its metabolite
27 psychosine. We have shown previously that the S1P receptor agonist fingolimod (FTY720)
28 attenuates psychosine-induced glial cell death and demyelination both in vitro and ex vivo
29 models. This data, together with a lack of therapies for Krabbe’s disease, prompted the current
30 preclinical study examining the effects of fingolimod in twitcher mice, a murine model of
31 Krabbe’s disease. Twitcher mice, both male and female, carrying a natural mutation in the
32 GALC gene were given fingolimod via drinking water (1mg/kg/day). The direct impact of
33 fingolimod administration was assessed via histochemical and biochemical analysis using
34 markers of myelin, astrocytes, microglia, neurons, globoid and immune cells. The effects of
35 fingolimod on twitching behaviour and lifespan was also demonstrated. Our results show that
36 treatment of twitcher mice with fingolimod significantly rescued myelin levels compared to
37 vehicle treated animals and also regulated astrocyte and microglial reactivity. Furthermore,
38 non-phosphorylated neurofilament levels were decreased indicating neuroprotective and
39 neurorestorative processes. These protective effects of fingolimod on twitcher mice brain
40 pathology, was reflected by an increased lifespan of fingolimod treated twitcher mice. These
41 in vivo findings corroborate initial in vitro studies and highlight the potential use of S1P
42 receptors as drug targets for Krabbe’s disease.
43
44 Keywords: Globoid cell leukodystrophy; Krabbe’s disease; FTY720; Fingolimod; myelination,
45 neurodevelopmental disease
2 | Page
46 Significance Statement
47 This study demonstrates that administration of the therapy, known as fingolimod, in a mouse
48 model of Krabbe’s disease (namely, the twitcher mouse model) significantly rescues myelin
49 levels. Further, this drug fingolimod, also regulates the reactivity of glial cells, astrocyte and
50 microglia, in this mouse model. These protective effects of fingolimod result in an increased
51 lifespan of twitcher mice.
52
53
3 | Page
54 Introduction
55 Krabbe’s disease (globoid cell leukodystrophy) is a devastating illness that is invariably fatal
56 within the first two years of life (Graziano and Cardile 2015). This disease has orphan status
57 affecting approximately 1:100,000 births, although the incidence varies in different
58 populations (Barczykowski et al. 2012). Krabbe’s disease is an inherited lipid storage disorder
59 resulting from oligodendrocyte cell death and subsequent loss of myelin. The disease is caused
60 by mutations in the galc gene encoding for galactosylceramidase (GALC) (Suzuki and Suzuki
61 1970). Mutations in GALC result in enzymatic dysfunction and a build-up of its two metabolites
62 galactosylceramide and the toxic galactolipid galactosylsphingosine (psychosine) (Suzuki
63 1998). Aggregations of the latter is particularly apparent in the white matter of the brain and
64 in sciatic nerves, where it has been shown to inhibit some critical cell processes resulting in
65 oligodendrocyte and Schwann cell apoptosis (Giri et al. 2008, Misslin et al. 2017). Pathological
66 features of Krabbe’s disease therefore include profound demyelination and almost complete
67 loss of oligodendrocytes in the white matter, accompanied by inflammatory mechanisms
68 including reactive astrocytosis and infiltration of numerous multinucleated phagocytes termed
69 ‘globoid cells’ (Suzuki 2003).
70
71 The clinical phenotype of Krabbe’s disease is classified based on the age of disease onset, with
72 the majority of cases affecting infants (Wenger et al. 2016). Infantile Krabbe’s disease typically
73 develops within the first six months postnatal with progressive rapid neurologic deterioration.
74 Hallmark symptoms of the classic infantile forms include irritability, hypertonic spasticity and
75 psychomotor stagnation, followed by rapid developmental decline, seizures and optic atrophy
76 (Graziano and Cardile 2015). Clinical manifestations thus suggest involvement of both the first
4 | Page
77 and second motor neurons, indicative of a systemic disorder affecting the central as well as
78 the peripheral nervous systems. To date there is no therapeutic cure for Krabbe’s disease. A
79 number of therapeutic strategies have been described, targeting various levels of the
80 pathomechanistic cascade in order to lower the psychosine load and reduce its neural toxicity
81 (Bongarzone et al. 2016). The current standard of care for patients with Krabbe’s disease is
82 limited to hematopoietic stem cell transplantations, derived from bone marrow or umbilical
83 cord blood (Escolar et al. 2005). While this treatment has been shown to slow disease
84 progression, it fails to correct peripheral neuropathy in infants (Escolar et al. 2005, Duffner et
85 al. 2009). With increasing evidence suggesting Krabbe’s disease to be a multi-modal illness,
86 which includes ongoing inflammatory and neuronal pathologies, a combination of therapies
87 targeting these processes may prove more promising.
88
89 Previously, we have shown that psychosine causes human and mouse astrocyte toxicity in
90 culture and demyelination in mouse organotypic slice cultures, effects which were attenuated
91 by sphingosine 1-phosphate receptor (S1PR) agonists fingolimod and siponimod (O'Sullivan
92 and Dev 2015, O'Sullivan et al. 2016). S1PRs are G-protein coupled and expressed in many cell
93 types including immune, cardiovascular and central nervous systems (Dev et al. 2008). The
94 drug fingolimod targets all five S1PR subtypes, apart from S1PR2 and is marketed as the first
95 oral therapy for relapsing-remitting multiple sclerosis (Kappos et al. 2015). Fingolimod is
96 described to work by internalising S1PRs in T-cells, thus limiting their egress from lymph nodes
97 and dampening inflammation in multiple sclerosis (Dev et al. 2008). Furthermore, it has been
98 extensively demonstrated that S1PRs regulate neuronal and glial cell function. Briefly, in glial
99 cells, S1PRs play a role in oligodendrocyte differentiation, survival and myelination state,
100 astrocyte cell migration, survival and cell signalling, and microglia reactivity and pro-
5 | Page
101 inflammatory cytokine release (Miron et al. 2008, Miron et al. 2010, Sheridan and Dev 2014,
102 O'Sullivan and Dev 2015, O'Sullivan et al. 2016).
103
104 Given fingolimod may have potential to alter both inflammatory and neuronal dysfunction in
105 both the brain and periphery, and having previously shown that S1PR agonists attenuate
106 psychosine-induced cell death of astrocytes and demyelination in vitro, the current study
107 aimed to examine in vivo effects of fingolimod in twitcher mice, the murine model of Krabbe’s
108 disease.
6 | Page
109 Materials and Methods
110 Animals and experimental design
111 A breeding colony of heterozygous twitcher mice was established in a pathogen free
112 environment in the Comparative Medicine Unit Trinity College Dublin using mice obtained from
113 the Jackson Laboratory and maintained on a C57BL/6j genetic background (C57BL/6J-twi with
114 C57BL/6J-twi). Homozygous and heterozygous animals were identified using genotyping.
115 Heterozygous animals were only used for breeding purposes, and not used in the study.
116 Fingolimod (SML0700, Sigma) was administered to both male and female wildtype and
117 homozygous twitcher mice via drinking water (at a concentration of 6.25 ђg/mL), that would
118 result in a calculated final dose of 1mg/kg/day.
119 Behavioural analysis
120 All experiments were conducted in accordance with EU guidelines and authorised by the
121 Health Products Regulatory Authority. All animals were subjected to a behavioural analysis
122 beginning at 21 days of age (postnatal day 21, PND21), on the day that treatment with
123 fingolimod started. Behavioural observations were taken daily up to 3 times per day and body
124 weight was measured at each behaviour testing time point. An observational scale was used
125 to classify both the frequency and the severity of twitching during disease progression. For
126 twitching frequency, a modified scoring system of a previously published protocol was used
127 (Wicks et al. 2011). Very mild and fine twitching similar to post-weaning symptoms were scored
128 as Score 1. Mild, intermittent fine twitching was given a Score 2, while constant fine twitching
129 corresponded to a Score 3. Constant moderate twitching of body and head was given a Score
130 4, and followed by euthanisation if Scored 4 during four independent observations. Constant
7 | Page
131 and severe trembling and uncontrollable twitching was the highest score, and followed
132 immediately by humane euthanisation. The locomotor deficits arising as a consequence of
133 demyelination were quantified using a standardised manual scoring system that has been
134 previously published and accepted for scoring mobility in animal models of Multiple Sclerosis
135 (Beeton et al. 2007). A modified version of this scoring system as per other previous
136 publications using twitcher mice was applied, with a minimum score of one indicating absence
137 of disease, and a maximum score of 4. Due to overt phenotypic emergence in twitcher mice,
138 the genotype of animals used in this study became evident during the study, which limited
139 somewhat genotype-based blinding. Water bottles were filled with drug by the experimenter
140 and thus no blinding was included at the level of drug treatment. All data was randomized and
141 analysis was performed with data blinded to both genotype and drug treatment.
142
143 Processing of tissue for immunohistochemistry and immunoblotting
144 Mice were euthanized by placing in a CO2 chamber. The spleen and kidneys were removed and
145 placed in RIPA buffer and animals were transcardially perfused with 20mL of ice-cold PBS. Once
146 perfusion was complete, animals were decapitated and brains were removed and split into two
147 hemispheres. One hemisphere was further separated into cerebellum and cortex. Together
148 with spinal cord and sciatic nerves, cerebellum and cortex were stored in RIPA buffer
149 supplemented with protease inhibitors at minus 80°C until processed for biochemical analysis.
150 The remaining hemisphere was post-fixed in 4% paraformaldehyde overnight (4°C) before
151 being cryoprotected in 30% sucrose solution (4°C). Brains were then snap-frozen in isopentane
152 on dry ice and embedded in optimal cutting temperature compound (O.C.T.) Parasagittal
153 cryosections of 12ђm thickness were cut and used for immunohistochemistry.
8 | Page
154 Histology and immunofluorescence
155 Cerebellar cryosections were equilibrated to room temperature, and then rehydrated with
156 PBS. Sections were permeabilized using 0.1% triton X-100 (Tx, Sigma, T9284) in PBS to reduce
157 non-specific binding and blocked with 10% bovine serum albumin (BSA, Sigma, 10735086001)
158 for 2 hours at room temperature. Primary antibody incubations were conducted over night at
159 4°C in PBS containing 2% BSA and 0.5% Tx. Primary antibodies used were anti-vimentin (1:1000
160 dilution, Santa Cruz; sc-373717), anti-GFAP (1:1000 dilution, Abcam, ab7260), anti-NFH
161 (1;1000 dilution, Millipore; MAB5539), anti-SMI-32 (1:1000 dilution, Millipore), anti-MBP
162 (1:1000 dilution, Abcam; ab40390), anti-degraded MBP, dMBP (1:1000 dilution, Millipore;
163 AB5864), anti-MOG (1:1000 dilution, Millipore, MAB5680), anti-Iba1 (1:1000 dilution; Abcam;
164 ab5076) and anti-Calbindin (1:1000 dilution, Abcam, ab108404). Secondary antibody
165 incubations were carried out overnight at 4ȗC in PBS supplemented with 2% BSA and 0.05% Tx.
166 Secondary antibodies used included goat anti-rabbit Alexa 488 (1:1000, Invitrogen, A11008),
167 goat anti-chicken Alexa 633 (1:1000, Invitrogen, A21103), and anti-mouse DyLight 549 (1:1000,
168 Invitrogen). A counterstain with Hoechst 33342 (1:10,000, Thermo Scientific, 62249) labelling
169 the nucleus was performed at the end of the immunofluorescence protocol. Slices were
170 mounted and coverslipped on microscope slides and stored in the dark before being imaged.
171 For Periodic Acid Schiff (PAS, Fischer Scientific, 88016-88017) and Luxol Fast Blue (LFB, Acros
172 organics, 10348250) staining, the tissues were equilibrated to room temperature and
173 rehydrated with PBS. LFB and PAS staining was performed according to standard histology
174 procedures.
175 Light and fluorescence microscopy
9 | Page
176 Data acquisition and quantification was similar to our previous studies (Sheridan and Dev
177 2014). Confocal images captured for quantitative measurement of immunofluorescent staining
178 were 12 bit.lif files of 1024x1024 or 2048x2048 pixel resolution and were randomly acquired
179 throughout the cerebellum. They were captured using a Leica Sp8 scanning confocal with 10x
180 and 20x objectives. With 5 cerebellar slices per slide, approximately 10-12 images were taken
181 per slide to cover most of the total area of the cerebellum. Between 6 and 8 slides were used
182 per treatment group making a minimum of 60 images analysed per treatment group. Image
183 acquisition settings were kept the same across different treatments per experiment. Image
184 analysis was conducted using the software ImageJ (https://imagej.nih.gov/ij/). Regions of
185 Interest (ROI) were manually selected of each image and fluorescence intensity was averaged.
186 Fluorescence intensity results are shown in arbitrary fluorescence units. For Iba1
187 quantification, we used z-stacks to compute 3D models of Iba1-positive microglia with the
188 software Imaris (Bitplane) rather than measuring fluorescence intensity. The surface of those
189 3D models was then used to identify area and volume taken up by microglia, with the
190 hypothesis that the amoeboid and reactive state of microglia in twitcher mice will be reflected
191 in higher volume measurements.
192 Immunoblotting
193 Samples from cerebellum, spinal cord and sciatic nerve were homogenized and sonicated in
194 RIPA buffer containing protease inhibitors (cOmplete, Roche, 11697498001), centrifuged at
195 14,000rpm and supernatant was collected. For western blotting, samples were denatured and
196 electrophoresis carried out on 12% SDS-polyacrylamide gels. Electrophoresis was followed by
197 a semi-dry transfer to a PVDF membrane (Millipore, IPVH00010), which was then blocked in
198 5% Marvel in 0.05% PBS/Tween 20 at room temperature. Incubation with primary antibody
10 | Page
199 was performed over night at 4°C. Primary antibodies used were rabbit anti-MBP (1:2000,
200 Abcam, ab40390), mouse anti-MOG (1:1000, Millipore, MAB5690), rabbit anti-Olig2 (1:2000,
201 Abcam, ab109186), mouse anti-Vimentin (1:1000, Santa Cruz, sc-373717) and chicken anti-
202 GFAP (1:5000, Abcam, ab7260). The membranes were washed and incubated with secondary
203 goat anti-rabbit (VWR, NA934, 1:5000) or goat anti-mouse (Biosciences, A16066, 1:5000) HRP-
204 conjugated antibodies for 2 hours at room temperature. Membranes were developed using
205 chemiluminescent HRP substrate (VWR, PIER34580) and images were acquired and analysed
206 on a C-Digit blot scanner with Image studio 4.0 (LI-COR®).
207 Statistical analysis
208 Experimental data was analysed and graphically represented using GraphPad Prism 8.0
209 Software package (GraphPAD Software, Inc.). The normality of the data was determines using
210 the Shapiro-Wilk test. The difference in the lifespan between twitcher and wildtype mice was
211 examined by Kaplan-Meier log-rank analysis. Discrepancies in weight gain, twitching severity
212 and mobility impairments were analysed using the two-way ANOVA with Bonferroni post-hoc
213 test for multiple comparisons. Mean fluorescence intensity as measured with ImageJ software
214 was used as an arbitrary unit of measure. Raw data sets were normalised and presented as
215 percentages of the control group average. For that purpose, each arbitrary value in the
216 experimental group was divided by the mean of the control group and multiplied by 100, in
217 order to be expressed as a percentage value of the control average. Statistical evaluation of
218 experimental data was performed with the two-way ANOVA followed by Bonferroni post-hoc
219 test, with p<0.05 as the minimum level of significance. Graphical data is represented as mean
220 ±SEM.
11 | Page
221 Results
222 Fingolimod regulates levels of myelin in twitcher animals
223 We have demonstrated that fingolimod promotes remyelination and inhibits demyelination in
224 psychosine treated cerebellar slice cultures, as well as in lysophosphatidylcholine (LPC), H2O2
225 and splenocyte treated cerebellar slice cultures (Pritchard et al. 2014, Sheridan and Dev 2014,
226 O'Sullivan and Dev 2015). Here, we examined myelin status in vehicle- and fingolimod-treated
227 twitcher mice, using myelin basic protein (MBP) and myelin oligodendrocyte protein (MOG),
228 both of which constitute late markers of oligodendrocyte maturation and are involved in the
229 final stages of myelin compaction. As expected, fingolimod did not alter the amount of myelin
230 in wildtype littermates under non-pathological conditions (FFigure 1), an observation that
231 coincides with previous findings (Miron et al. 2010). As anticipated, MBP and MOG
232 immunostaining in vehicle treated twitcher mice was significantly reduced compared to their
233 wild-type littermates (MBP: 100±11.23 vs 47.51±3.72,**p=0.0075, n=10, df=33) (MOG:
234 100±9.9 vs 59.97±3.4, **p=0.0036, n=8, df=26) (FFigure 1). Importantly, we found the
235 administration of fingolimod in twitcher mice increased significantly MBP expression
236 compared to vehicle-treated animals (47.51±3.72 vs 94.66±7.78, *p=0.04, n=7-10, df=33)
237 (FFigure 1), with no significant effect on MOG expression (59.97±3.4 vs 82.75±3.67, n.s.
238 p=0.248, n=7-8, df=26) (FFigure 1).
239 The effects of fingolimod on the total levels of MOG and MBP in wildtype and twitcher mice
240 were also examined using homogenized cerebellum (FFigure 2A,B) and spinal cord (FFigure 3A,B)
241 tissue by Western blotting. While, the levels of MBP in the cerebellum and MOG in the spinal
242 cord were reduced in twitcher mice compared to vehicle controls, fingolimod did not alter the
243 total expression levels of these markers (FFigures 2A,B, 3A,B). The expression levels of Olig2 was
12 | Page
244 also examined, which was significantly decreased in twitcher mice compared to wildtype
245 littermates (FFigures 2C, 3C). In this case, fingolimod enhanced significantly the levels of Olig2,
246 in twitcher mice compared to vehicle treated in the cerebellum (Olig2: 18.209±6.11 vs
247 50.684±11.288, *p=0.0386, n=9-10, df=35)) (FFigure 2C).
248 To further investigate the effects of fingolimod we also examined levels of myelin debris, by
249 using an antibody that recognises degraded MBP. As expected, the data showed a significant
250 increase in of myelin debris in twitcher mice compared with wildtype controls (100±7.3 vs
251 171±13, **p=0.0033, n=4, df=10) (FFigure 4). We note here, that the administration of
252 fingolimod did not alter levels of myelin debris in wildtype or twitcher mice (171±13 vs
253 171.3±5.324, p>0.9990, n=4, df=10) (FFigure 4). Overall, the data indicates that fingolimod
254 rescues demyelination in the brains of twitcher animals via a mechanism likely involving an
255 increase in oligodendrocyte cells expressing MBP and Olig2, with little effect on total levels of
256 myelin or on myelin debris.
257
258 Fingolimod restores Bergmann glial cell expression and attenuates astrocyte
259 activation
260 Psychosine induces death of human and mouse astrocytes in vitro that is attenuated by pre-
261 treatment with fingolimod (O'Sullivan and Dev 2015). Here, we examined astrocytes using the
262 type III intermediate filament astrocyte marker vimentin that is involved in cytoskeleton
263 formation in astrocytes ( Figure 5A), as well as examining the glial fibrillary acidic protein
264 (GFAP), which is enhanced following astrocyte activation ( Figure 5B). We differentiated
265 between Bergmann glial cells in the molecular layer (ML) and fibrous astrocytes located in the
13 | Page
266 white matter (WM), as depicted by cresyl violet (Figure 5C). A significant decrease in vimentin
267 fluorescence occurred between vehicle-treated wildtype and twitcher mice in the molecular
268 layer of the cerebellum (ML: 100±14.86 vs 50.79±4.42, **p=0.0131, n=5, df=14) (FFigure 5D).
269 In agreement with our previous in vitro data, administration of fingolimod promoted
270 significantly the expression of vimentin in the molecular layer of twitcher mice, as compared
271 to their vehicle treated littermates (ML:50.79±4.42 vs 90.32±11.3, *p=0.0394, n=5, df=14)
272 (FFigure 3D). Conversely, in the WM of twitcher mice, there was an increased vimentin
273 fluorescence in vehicle treated twitcher mice compared to control (WM: 100±27.25 vs
274 239.83±27.15, **p=0.0067, n=5, df=16), which was further increased in fingolimod-treated
275 twitcher animals (WM: 239.83±27.15 vs 382.3±24.18, **p=0.0033, n=5-6, df=16) (FFigure 5D).
276 GFAP immunostaining showed no significant differences in the ML or WM layers of the
277 cerebellum of untreated or fingolimod-treated twitcher mice compared to healthy controls
278 (WM: 100±15.87 vs 162.6±20.75, n.s p=0.0699, n=5, df=16; ML: 100±15.82 vs 136.644±29.391,
279 n.s p>0.9999, n=4-6, df=18) (Figure 3E). Bergmann glia are generally aligned to Purkinje cell
280 somata in ascending processes spanning the molecular layer. We noted this organised
281 astrocyte scaffold was disrupted in twitcher mice. Notably, however, higher magnified images
282 of vimentin and GFAP immunostaning did not show any overt qualitative differences in fibrous
283 astrocyte morphologies. To examine if fingolimod altered the total levels of vimentin and GFAP,
284 Western blotting was performed using homogenized cerebellum (FFigure 2D,E) and spinal cord
285 (FFigure 3D,E). While the levels of vimentin and GFAP increased in twitcher mice compared to
286 vehicle controls, fingolimod-treatment did not alter total levels of these markers (Figures 2D,E,
287 3D,E). Overall, these data suggest that astrocytes have an altered function in the cerebellum
288 of twitcher mice, and this is modulated by fingolimod.
289
14 | Page
290 Fingolimod decreases microglia levels in the cerebellar white matter of twitcher
291 mice
292 The CNS of patients with KD, as well as, that of animal models of KD exhibit robust
293 neuroinflammation with microglial activation and the presence of phagocytic multinucleated
294 globoid cells (Potter and Petryniak 2016). Fingolimod possesses potent anti-inflammatory
295 effects in various animal models of CNS injury and neurodegeneration (Aytan et al. 2016,
296 Martinez and Peplow 2018). Furthermore, fingolimod shifts LPS-activated microglia towards a
297 neuroprotective microglial phenotype (Cipriani et al. 2015). We therefore investigated the
298 effect of fingolimod treatment as a neuroinflammatory modulator in twitcher mice. In our
299 hands, and surprisingly, we have not observed psychosine to alter the expression of Iba1
300 (ionized calcium binding adapter molecule 1) in organotypic cerebellar slice cultures (O'Sullivan
301 and Dev 2015), although we note Iba1 is not a specific marker of altered microglia reactivity.
302 In contrast to this, and in line with our expectations, here we found a significant increase in
303 Iba1 immunostaining in vehicle-treated twitcher mice, as compared to their wildtype
304 littermates (2.675±0.5 vs 120.9±16.02, p<0.0001, n=5, df=13) (FFigure 6A-C). More importantly,
305 fingolimod significantly decreased Iba1 expression in twitcher mice compared to vehicle
306 treated controls (120.9±16 vs 32.93±3.5, p<0.0001, n=5, df=13) (FFigure 6A-C). Higher
307 magnified images of Iba1 immunostaining showed amoeboid morphology in twitcher mice
308 compared to wildtype littermates, which appeared less so in twitcher mice administrated with
309 fingolimod, although these microglia also still appeared very much of an amoeboid structure
310 (FFigure 6B).
311
15 | Page
312 Multinucleated phagocytes are a histopathological hallmark of KD and have been shown to
313 accumulate large amounts filamentous inclusions that stain positively for periodic acid-Schiff.
314 The distribution of these globoid cells in the cerebellar white matter, of treated and untreated
315 twitcher mice, was examined also by staining with luxol fast blue and periodic acid-Schiff. Our
316 data showed increased levels of periodic acid-Schiff staining in vehicle treated twitcher mice
317 (1.48 ±0.25x100ђm2), which was reduced significantly in fingolimod-treated animals (0.75
318 ±0.05x100ђm2) (FFigure 6D-F). These observations demonstrate increased Iba1 expression and
319 increased globoid cell expression in twitcher mice, which is attenuated significantly by
320 fingolimod.
321
322 Effects of Fingolimod on phosphorylation state of neurofilaments and purkinje
323 cells
324 Neurofilaments are neuron-specific cytoskeletal proteins essential for the development and
325 maintenance of neurons and their processes. They form a structurally and genetically related
326 protein subunits, which are thought to be associated with the level of myelination and to be
327 related to the fast conduction of axons (Yuan et al. 2012). Previous studies have shown that
328 non-phosphorylated neurofilament heavy (NFH) epitopes are expressed in neuronal cell bodies
329 and purkinje cells under physiological conditions (Demilly et al. 2011). Increased expression of
330 non-phosphorylated NFH in axonal tracts, however, has been associated with impaired axonal
331 conduction, demyelination and neuronal damage (Misslin et al. 2017). In order to examine NFH
332 phosphorylation state in twitcher mice, cerebellar slices were stained for NFH and SMI-32, a
333 marker known to label non-phosphorylated NFH. A qualitative observation showed increased
334 expression of SMI32 in twitcher animal, as expected, which was decreased in fingolimod-
335 treated twitcher mice, as reflected by a change in mean quantification of SMI32 fluorescence
16 | Page
336 intensity. We note, however, a level of variance in the expression of SMI32 across animals, and
337 more so in the twitcher animals compared to wildtype controls, that resulted in non-significant
338 statistical analysis (Figure 7A-D). Cerebellar purkinje cells are pivotal for motor coordination
339 and their dysfunction generally leads to ataxia (Hoxha et al. 2018). The ataxic tremors and hind
340 limb paralysis observed in human and murine KD suggests the pathological involvement of
341 purkinje cells within cerebellar degeneration. Given that little is known regarding the pathology
342 of the purkinje cell layer in KD, we investigated the number of purkinje cells in this layer by
343 immunostaining with Calbindin. No overall difference in the number of Calbindin-positive
344 purkinje cells was observed (Figure 7E) in twitcher mice compared with wildtype littermates
345 and compared to twitcher mice treated with fingolimod (Figure 7E-F). Structural changes were
346 observed, including the presence of heterotopic layers of purkinje cells of fingolimod- and
347 vehicle-treated twitcher mice, with their soma displaced in the molecular layer (Figure 7G).
348 This data suggests that neuronal damage in twitcher mice is not necessarily driven by purkinje
349 cell loss, but rather by their abnormal architectural integrity.
350
351 Fingolimod enhances lifespan of twitcher mice
352 The administration of fingolimod from postnatal day 21 (PND) onwards (FFigure 8A), in twitcher
353 mice, significantly improved their mobility (FFigure 8B) and twitching severity (FFigure 8C), which
354 we noted had relapsing/remitting trends. Fingolimod also positively affected body weight, with
355 significant improvement between PND25 and PND30 (FFigure 8D), although this weight gain
356 plateaued at PND30 and finally converged towards vehicle treated animals. Importantly,
357 Kaplan-Meier survival curves demonstrated fingolimod to modestly but significantly increase
358 lifespan of twitcher mice versus vehicle treated littermates (FFigure 8E). We noted this effect
359 showed an approximate 10% increase in average lifespan of fingolimod (42 days) versus vehicle
17 | Page
360 (37 days) treated twitcher mice (FFigure 8E). While we had hoped for a more promising effect
361 on lifespan, this result reflects the well-known multi-organ involvement of leukodystrophies,
362 where neuroprotective effects of drugs such as fingolimod may become an important part of
363 a combination therapy that target both central and peripheral systems in these illnesses.
18 | Page
364 Discussion
365 Summary of Findings
366 The rapid and complete loss of myelin and myelin forming oligodendrocytes is a predominant
367 pathological feature of Krabbe’s disease (KD), and is responsible for slowed nerve conduction
368 observed in twitcher animals, a natural occurring animal model of KD (Davenport et al. 2011).
369 With studies reporting remyelinating effects of fingolimod in relapsing-remitting MS (Münzel
370 and Williams 2013), and based on our previous in vitro work demonstrating fingolimod-
371 mediated protective effects against astrocyte cell toxicity and demyelination, we anticipated
372 this drug to display efficacy in other demyelinating diseases, such as KD. Here, we investigated
373 the efficacy of fingolimod treatment on phenotypic and underlying and biochemical
374 abnormalities in twitcher mice, the murine model of KD. The present results enabled a direct
375 demonstration that fingolimod can rescue deficits seen at the pathological level, namely levels
376 of demyelination, as well as promote lifespan and attenuate phenotypic dysfunction in these
377 animals.
378 The regulation of oligodendrocytes by fingolimod
379 With fingolimod being able to cross the blood brain barrier and act on S1PRs within the CNS,
380 this drug has been postulated to enhance and stabilize myelination by direct action on
381 oligodendrocyte lineage cells through S1PR1,3,5 subtypes (Miron et al. 2010). Specifically,
382 mature oligodendrocytes express high levels of S1PR5, while oligodendrocyte precursor cells
383 (OPCs) show preferred expression of S1PR1 (Choi and Chun 2013). Moreover, remyelination
384 requires OPCs to proliferate, migrate to sites of demyelination, and finally to differentiate into
385 mature myelin-forming oligodendrocytes, for which S1PRs play a role (Miron et al. 2011). The
386 observed enhanced Olig2 expression in cerebellar samples, in conjunction with increased
19 | Page
387 expression of the myelination marker MBP-induced by fingolimod may suggest a remyelination
388 event that involves an increase in OPC numbers, although we note the caveat that Olig2 is not
389 expressed only in OPCs. Treatment with pFTY720 regulates also the cell survival and
390 differentiation in oligodendrocyte lineage cells in cultured rat oligodendrocytes and OPCs
391 (Jaillard et al. 2005, Jung et al. 2007). Additionally, S1P1 activation is involved in OPC
392 mitogenesis (Jung et al. 2007). In a cuprizone model of demyelination, fingolimod increases
393 the number of OPCs, without however promoting remyelination (Kim et al. 2011). Importantly,
394 pFTY720 modulates mature oligodendrocyte membrane dynamics and survival responses in
395 human oligodendrocyte cultures (Miron et al. 2008). Together these findings support the
396 possibility of an immediate protective effect of S1PR modulation on oligodendrocytes through
397 various signalling pathways.
398 Fingolimod regulates microglia in twitcher mice
399 The psychosine hypothesis was introduced in 1972 and attributes the underlying cellular and
400 biochemical pathology of KD to supraphysiological aggregations of psychosine in the periphery
401 and CNS (Suzuki 1998). Emerging evidence complements this longstanding hypothesis by
402 ascribing early demise in KD to neuroinflammation rather than myelin loss (Nicaise et al. 2016).
403 In addition to fingolimod positively regulating the levels of myelin in twitcher mice, our current
404 study support the involvement of immunomodulation via glial cells. Activation of astrocytes
405 and microglia have been reported to play a compounding role in disease progression of KD
406 (Mohri et al. 2006, Ijichi et al. 2013). Microglia mediate inflammation through increased levels
407 of prostaglandin (PG) D2, while astrocytes increase PGD2 receptor expression (Mohri et al.
408 2006). This study reported suppressed astrogliosis and reduced demyelination when blocking
409 PGD2 synthase in twitcher mice (Mohri et al. 2006). The remyelination in fingolimod-treated
20 | Page
410 twitcher animals observed in our current study coincides with the reduction we observed in
411 the immunostaining of Iba1, albeit as surrogate marker of microglia reactivity. These findings
412 correlate with the hypothesis that regulation of S1PRs mediate enhanced levels of myelination
413 via a mechanism that involves immunomodulatory pathways and microglia (Ijichi et al. 2013).
414 The S1P axis regulates astrocytes and astrogliosis in twitcher mice
415 A growing body of evidence in KD research, paints a complex picture of combined axonal,
416 neuronal, and myelination abnormalities in the CNS and periphery, accompanied by
417 exacerbating astrogliosis and neuroinflammatory events (Nicaise et al. 2016, Won et al. 2016).
418 Dysfunctional astrocytes may be both a contributing and a combating factor to disease
419 progression of various neurological disorders (Molofsky et al. 2012). Astrocytes are implicated
420 in the pathogenesis of KD, where the astrocytic expression of matrix metalloproteinase (MMP)-
421 3 is elevated at symptomatic onset in twitcher mice, continues to rise with disease progression
422 and potentially induces the formation of multinucleated globoid cells and activated phagocytes
423 (Ijichi et al. 2013, Claycomb et al. 2014). Such findings coincide with results from our current
424 study, where immunofluorescence in twitcher mice shows altered expression of vimentin in
425 areas of demyelination, accompanied by PAS-positive globoid cells. Our data showed
426 Bergmann glial cells in the molecular layer of the cerebellum stained with vimentin was
427 reduced in twitcher mice and this was attenuated by treatment with fingolimod. In contrast,
428 the fibrous astrocytes populating the white matter showed increased expression of vimentin
429 in twitcher mice, which was further enhanced by treatment with fingolimod. Both GFAP and
430 vimentin are highly expressed in white matter astrocytes, but poorly label gray matter
431 astrocytes (Molofsky et al. 2012). The differential expression of these intermediate filaments
432 suggests that fibrous astrocytes serve a specialised function with regards to structure among
21 | Page
433 myelinated fibres (Lundgaard et al. 2014). We note that in our study, we did not observe
434 statistical differences with GFAP staining between wild type and twitcher mice treated with or
435 without fingolimod. This could be explained by the fact that GFAP is expressed late in
436 development of fibrous astrocytes, and is therefore not an adequate marker to investigate
437 changes during early developmental stages (Lundgaard et al. 2014). Supporting the role of
438 S1PRs in astrocytes, pFTY720 attenuates psychosine-induced apoptosis in human astrocyte
439 cultures and pro-inflammatory cytokines in murine astrocyte cultures (O'Sullivan and Dev
440 2015). The selective deletion of S1PR1 from astrocytes also reduces EAE-related phenotypic
441 severity and histological abnormalities, including demyelination and astrogliosis (Choi et al.
442 2011). These previous studies together with our own, could be interpreted as a pFTY720-
443 induced phenotypic change in astrocytes.
444 Regulation of neuronal dysfunction through S1P modulation
445 Beside affecting oligodendrocytes, psychosine also impairs axonal transport (Cantuti-
446 Castelvetri et al. 2012). Neurons are not known to synthesise significant quantities of
447 psychosine, which is likely transferred from glial cells, resulting in neuronal damage and
448 impaired axonal conduction. Psychosine accumulation increases PP1 and PP2A phosphatases,
449 inducing neurofilament dephosphorylation and inhibiting axonal transport (Cantuti-Castelvetri
450 et al. 2012). Axonopathy precedes demyelination and neuronal apoptosis in twitcher mice,
451 with neuronal damage occurring at later stages of disease progression (Castelvetri et al. 2011,
452 Smith et al. 2011). Importantly, pFTY720 has protective effects on neuronal function and
453 axonal integrity, by reducing axonal damage following demyelination, for example, in a
454 cuprizone model or by regulating glutamateric transmission in EAE mice (Slowik et al. 2015).
455 These findings coincide with our observations using SMI32 as a marker for axonal damage,
22 | Page
456 where we observed a qualitative decrease in the expression of SMI32 in twitcher animals
457 administered with fingolimod, although we note this was not reflected by a quantitative
458 difference. Nevertheless, the current data sits in line with previous observations that suggest
459 S1PR modulation involves a concurrent preservation of axonal damage as well as restoration
460 of myelin levels.
461 S1P modulation and its impact on the cerebellar purkinje cell layer architecture
462 A striking pathological hallmark of KD encompasses cerebellar ataxia, found in 57% of infantile
463 onset cases (Debs et al., 2013). As disease progresses, twitcher mice display tremors, ataxia
464 and severe hind limb weakness. Such motor deterioration suggests involvement of cerebellar
465 degeneration in addition to pyramidal tract dysfunctions. Our results displayed structural
466 changes in the cerebellum of twitcher mice that appear to be characterised by a pruning of
467 dendritic arborisation and heterotopic layers of purkinje cells. These findings are in agreement
468 with other studies showing exacerbations in neuronal vulnerability and demyelination in the
469 internal granular, purkinje and molecular layers in cerebella of twitcher mice (Smith et al.
470 2014). Heterotopic purkinje cells are typically observed in neurodevelopmental and
471 neurodegenerative diseases, such as spinocerebellar ataxia type 1/6 or Essential Tremor, and
472 are defined by a mislocalised cell body in the molecular layer (Kuo et al. 2011). In the normal
473 cerebellum, purkinje cells form a monolayer between molecular and granular layers, with their
474 proper anatomical location as requirement for physiological cerebellar functioning. In a disease
475 characterised by severe cerebellar degeneration, such as KD, it is thus possible that aberrant
476 Purkinje cell layer architecture contributes to ataxic symptoms.
477 Combination approaches for treatment of KD.
23 | Page
478 Our results show that fingolimod treatment increased lifespan of twitcher mice by up to 5 days
479 and intermittently attenuated phenotypic abnormalities. While disappointing, these modest
480 improvements may be explained by ongoing disease progression in the peripheral nervous
481 system and other visceral organs. In supplementary studies, we observed little to no
482 improvements in sciatic nerve demyelination in twitcher mice administered with fingolimod
483 (Figure 9). Furthermore, we noted a consistent decrease in size of visceral organs such as
484 spleen and kidneys of twitcher mice, compared to healthy controls (data not shown). These
485 observations imply that fingolimod-mediated improvements in the central nervous system are
486 not sufficient to correct for the multiple organ failure, weight loss and peripheral degeneration.
487 Weight measurements of fingolimod treated twitcher animals ultimately converges back to
488 that of untreated littermates, and animals eventually develop tremors, flaccid paralysis of the
489 hindlimbs that likely results from failed correction of peripheral nerve function and
490 demyelination. In support of this, failure to correct peripheral dysfunction in KD, is considered
491 a challenge in infants treated with pre-symptomatic hematopoiteitc stem cell transplantation
492 (Escolar et al. 2005, Wenger et al. 2016). Thus, our observations, as well as previous studies,
493 highlight the need to target both central and peripheral systems in order to provide a more
494 effective therapy in KD (Lin et al. 2005, Rafi et al. 2012).
495
496
24 | Page
497 Figure Legends
498 Figure 1. Levels of MBP in twitcher animals are rescued by fingolimod. 499 (A) Cerebellar slices of vehicle and FTY720-treated twitcher and wildtype mice were processed for 500 immunohistochemistry. Representative images display MOG (yellow), MBP (green) and NFH (red) 501 immunostaining under treatment conditions are indicated. Confocal images were captured at 10x 502 magnification (scale bar 200ђm) with focus on cerebellar lobes. (B) Confocal images were captured at 20x 503 magnification (scale bar 100ђm) with focus on inner white matter and lobe VII. (C) Graphs illustrating 504 changes in MOG and MBP fluorescence post treatment. Differences between treatment groups were 505 analysed using a two-way ANOVA followed by a Bonferroni correction for multiple comparisons (*p<0.05; 506 **p<0.01; ***p<0.001, n=7-10 animals). Graphical data is represented as mean ±SEM. 507 508 Figure 2. Changes in protein expression levels in the cerebellum of twitcher mice. 509 Western blots on cerebellar tissue. (A) Total MOG remained at similar levels across all four treatment 510 groups. (B) Data showing significant decrease in total MBP levels in twitcher mice compared to healthy 511 littermates, with no significant effect of fingolimod. (C) Olig2 was reduced significantly in untreated 512 twitcher mice, which was increased significantly with fingolimod treatment (*p<0.05, ****p<0.0001, 513 n=10). Both (D) vimentin and (E) GFAP astrocyte markers were significantly increased in both treated and 514 untreated twitcher mice. For vimentin, we noted the appearance of three independent bands, which when 515 individually analysed followed a similar pattern as the combined total levels shown in the figure. Graphical 516 data is represented as mean ±SEM. 517 518 Figure 3. Changes in protein expression levels in the spinal cord of twitcher mice. 519 Western blots on spinal cord tissue. Myelination as measured by (A) MBP and (B) MOG was significantly 520 decreased in untreated twitcher mice, which was not altered with fingolimod treatment. (C) Olig2 was 521 significantly reduced in untreated twitcher mice, with no significant effect in by fingolimod administration 522 in twitcher mice (*p<0.05, n=7). As observed in the cerebellum, the astrocyte markers (D) vimentin and (E) 523 GFAP were significantly increased in both treated and untreated twitcher mice. Fingolimod did not 524 significantly change protein levels of vimentin or GFAP in treated twitcher mice. Graphical data is 525 represented as mean ±SEM. 526 527 Figure 4. Impaired myelin debris clearance in the cerebellar white matter. 528 (A) Representative images showing degraded MBP (dMBP; purple), with the nuclear stain DAPI (blue) and 529 MOG (yellow) (scale bar 100 ђm). (B) Images were taken at 20x magnification, in the region of the 530 cerebellar white matter as depicted in the diagram (boxed region). (C) Graph illustrating changes in dMBP 531 across treatment groups, with high levels of myelin debris in both treated and untreated Twitcher mice
25 | Page
532 compared to their healthy littermates (**p<0.01; two-way ANOVA with Bonferroni’s post hoc test; n=4 533 animals, df = 10). 534 535 Figure 5. Fingolimod restores astrocyte reactivity in the cerebellum of twitcher mice. 536 Representative images displaying (A) vimentin (yellow) and (B) GFAP (red) in the cerebellum of wildtype 537 and twitcher mice under treatment conditions indicated (scale bar 200ђm). Magnified images (of twitcher 538 mice) are shown, (scale bar 50ђm). (C) Cresyl stain showing regions of interest analysed, namely, molecular 539 layer (ML) and white matter (WM). Graphs illustrating (D) vimentin and (E) GFAP expression within the ML 540 (upper graphs) and WM (lower graphs) of the cerebellum. Data presented as mean and median ±SEM. 541 Differences between groups were analysed using a two-way ANOVA followed by a Bonferroni correction 542 for multiple comparisons (*p<0.05; **p<0.01; ***p<0.001, n=5-6 animals). 543 544 Figure 6. Fingolimod attenuates Iba1 expression in the cerebellum of twitcher mice. 545 Representative images of (A) Iba1 (green) and NFH (red) immunostaining. Iba1 was significantly increased 546 in untreated twitcher mice, with accumulations centred around the white matter. (B) Magnified images 547 created using IMARIS software are shown, (scale bar 20ђm). (C) Graph illustrate Iba1-positive areas 548 expressed as mm2. Data presented as mean ±SEM. Differences between treatment groups were analysed 549 using a two-way ANOVA followed by a Bonferroni correction for multiple comparisons (n=4-5 animals). 550 Scale bar 200ђm. (D) Low magnification representative image of PAS stained globoid cells (purple) 551 accumulating along white matter tracts (blue) (E) High magnification representative images of PAS stain, 552 from region indicated in dotted box in (E) (scale bar 200ђm). (F) Graph illustrates globoid cell surface per 553 10ђm2 in twitcher mice treated with vehicle or fingolimod. Differences analysed with an unpaired t-test, 554 p<0.01 (n=4 animals). Graphical data is represented as mean ±SEM. 555 556 Figure 7. Fingolimod did not alter axonal damage or Purkinje cell layer architecture in twitcher mice. 557 (A) Representative images of SMI32 immunostaining. Confocal z-stacks were captured at 20x magnification 558 (scale bars 200ђm). (B-D) Graphical data showing SMI32 fluorescence intensity measurements, which were 559 taken across the different layers of the cerebellar cortex (n=6-7 animals). (E) Graphical data showing 560 Purkinje cell count did not change across treatment groups (n=5 animals). Data represented as mean ±SEM. 561 (F) Low magnification representative image of Calbindin immunostaining of Purkinje cells. High 562 magnification representative images of Calbindin immunostaining, from region indicated in dotted box. 563 Confocal z-stacks were captured at 20x magnification (scale bars 200ђm). (G) Representative images 564 showing presence of heterotopic layers of Purkinje cells in vehicle- and FTY720-treated twitcher mice and 565 to a certain degree in healthy wildtype animals (scale bars 100ђm). 566
567 Figure 8. Fingolimod enhances the lifespan of twitcher mice.
26 | Page
568 (A) Schematic representation of the study design. Fingolimod-treatment commenced at PND21 at a 569 concentration of 1mg/kg/day with daily behaviour monitoring. (B) Mobility, (C) twitching and (D) body 570 weight were assessed statistically using a two-way ANOVA followed by a Bonferroni adjustment for 571 multiple comparisons (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, n=11-15 animals). (E) The Kaplan- 572 Meier survival curve illustrates the lifespan of fingolimod treated versus untreated twitcher mice. A log- 573 rank Mantel-Cox test reveals that both treatment groups were significantly different in their survival 574 distribution (****p<0.0001, twitcher mice fingolimod treated versus untreated, n=20-33 animals), with an 575 increased average life span of FTY720-treated twitcher mice when compared to their vehicle-treated 576 twitcher littermates. Graphical data is represented as mean ±SEM. 577 578 Figure 9. Myelin Basic Protein changes in the sciatic nerve of twitcher mice. 579 MBP was used as a surrogate marker for measuring myelin levels in the sciatic nerves. As expected, 580 there was a significant decrease of MBP levels in untreated twitcher mice (p=0.0068, n=3), which was 581 not altered by fingolimod administration (p=0.0058, n=3). These findings thus suggest that fingolimod 582 was not able to modulate peripheral nerve demyelination. Graphical data is represented as mean 583 ±SEM.
584
27 | Page
585 References
586 Aytan N, Choi J-K, Carreras I, Brinkmann V, Kowall NW, Jenkins BG and Dedeoglu A (2016). "Fingolimod 587 modulates multiple neuroinflammatory markers in a mouse model of Alzheimer’s disease." Scientific 588 Reports 6: 24939. 589 Barczykowski AL, Foss AH, Duffner PK, Yan L and Carter RL (2012). "Death rates in the U.S. due to Krabbe 590 disease and related leukodystrophy and lysosomal storage diseases." American Journal of Medical Genetics 591 Part A 158A(11): 2835-2842. 592 Beeton C, Garcia A and Chandy KG (2007). "Induction and clinical scoring of chronic-relapsing experimental 593 autoimmune encephalomyelitis." J Vis Exp 5. 594 Bongarzone ER, Escolar ML, Gray SJ, Kafri T, Vite CH and Sands MS (2016). "Insights into the Pathogenesis and 595 Treatment of Krabbe Disease." Pediatr Endocrinol Rev 13 Suppl 1: 689-696. 596 Cantuti-Castelvetri L, Zhu H, Givogri MI, Chidavaenzi RL, Lopez-Rosas A and Bongarzone ER (2012). "Psychosine 597 induces the dephosphorylation of neurofilaments by deregulation of PP1 and PP2A phosphatases." 598 Neurobiol Dis 46(2): 325-335. 599 Castelvetri LC, Givogri MI, Zhu H, Smith B, Lopez-Rosas A, Qiu X, van Breemen R and Bongarzone ER (2011). 600 "Axonopathy is a compounding factor in the pathogenesis of Krabbe disease." Acta Neuropathol 122(1): 35- 601 48. 602 Choi JW and Chun J (2013). "Lysophospholipids and their receptors in the central nervous system." Biochim 603 Biophys Acta 1831(1): 20-32. 604 Choi JW, Gardell SE, Herr DR, Rivera R, Lee CW, Noguchi K, Teo ST, Yung YC, Lu M, Kennedy G and Chun J 605 (2011). "FTY720 (fingolimod) efficacy in an animal model of multiple sclerosis requires astrocyte 606 sphingosine 1-phosphate receptor 1 (S1P1) modulation." Proc Natl Acad Sci U S A 108(2): 751-756. 607 Cipriani R, Chara JC, Rodríguez-Antigüedad A and Matute C (2015). "FTY720 attenuates excitotoxicity and 608 neuroinflammation." Journal of neuroinflammation 12: 86-86. 609 Claycomb KI, Johnson KM, Bongarzone ER and Crocker SJ (2014). "An in vitro model for the study of cellular 610 pathophysiology in globoid cell leukodystrophy." J Vis Exp(92): e51903. 611 Davenport A, Williamson P and Taylor R (2011). "Pathohysiology of Krabbe disease." Orbit 2: 1-20. 612 Demilly A, Reeber SL, Gebre SA and Sillitoe RV (2011). "Neurofilament heavy chain expression reveals a unique 613 parasagittal stripe topography in the mouse cerebellum." Cerebellum 10(3): 409-421. 614 Dev KK, Mullershausen F, Mattes H, Kuhn RR, Bilbe G, Hoyer D and Mir A (2008). "Brain sphingosine-1- 615 phosphate receptors: implication for FTY720 in the treatment of multiple sclerosis." Pharmacol Ther 616 117(1): 77-93. 617 Duffner PK, Caviness VS, Jr., Erbe RW, Patterson MC, Schultz KR, Wenger DA and Whitley C (2009). "The long- 618 term outcomes of presymptomatic infants transplanted for Krabbe disease: report of the workshop held on 619 July 11 and 12, 2008, Holiday Valley, New York." Genet Med 11(6): 450-454. 620 Escolar ML, Poe MD, Provenzale JM, Richards KC, Allison J, Wood S, Wenger DA, Pietryga D, Wall D, 621 Champagne M, Morse R, Krivit W and Kurtzberg J (2005). "Transplantation of umbilical-cord blood in babies 622 with infantile Krabbe's disease." N Engl J Med 352(20): 2069-2081. 623 Giri S, Khan M, Nath N, Singh I and Singh AK (2008). "The role of AMPK in psychosine mediated effects on 624 oligodendrocytes and astrocytes: implication for Krabbe disease." J Neurochem 105(5): 1820-1833. 625 Graziano A and Cardile V (2015). "History, genetic, and recent advances on Krabbe disease." Gene 555(1): 2-13. 626 Hoxha E, Balbo I, Miniaci MC and Tempia F (2018). "Purkinje Cell Signaling Deficits in Animal Models of Ataxia." 627 Frontiers in Synaptic Neuroscience 10(6). 628 Ijichi K, Brown GD, Moore CS, Lee JP, Winokur PN, Pagarigan R, Snyder EY, Bongarzone ER and Crocker SJ 629 (2013). "MMP-3 mediates psychosine-induced globoid cell formation: implications for leukodystrophy 630 pathology." Glia 61(5): 765-777. 631 Jaillard C, Harrison S, Stankoff B, Aigrot MS, Calver AR, Duddy G, Walsh FS, Pangalos MN, Arimura N, Kaibuchi 632 K, Zalc B and Lubetzki C (2005). "Edg8/S1P5: an oligodendroglial receptor with dual function on process 633 retraction and cell survival." J Neurosci 25(6): 1459-1469. 634 Jung CG, Kim HJ, Miron VE, Cook S, Kennedy TE, Foster CA, Antel JP and Soliven B (2007). "Functional 635 consequences of S1P receptor modulation in rat oligodendroglial lineage cells." Glia 55(16): 1656-1667. 636 Kappos L, O'Connor P, Radue EW, Polman C, Hohlfeld R, Selmaj K, Ritter S, Schlosshauer R, von Rosenstiel P, 637 Zhang-Auberson L and Francis G (2015). "Long-term effects of fingolimod in multiple sclerosis: the 638 randomized FREEDOMS extension trial." Neurology 84(15): 1582-1591.
28 | Page
639 Kim HJ, Miron VE, Dukala D, Proia RL, Ludwin SK, Traka M, Antel JP and Soliven B (2011). "Neurobiological 640 effects of sphingosine 1-phosphate receptor modulation in the cuprizone model." Faseb j 25(5): 1509-1518. 641 Kuo S-H, Erickson-Davis C, Gillman A, Faust PL, Vonsattel J-PG and Louis ED (2011). "Increased number of 642 heterotopic Purkinje cells in essential tremor." Journal of neurology, neurosurgery, and psychiatry 82(9): 643 1038-1040. 644 Lin D, Fantz CR, Levy B, Rafi MA, Vogler C, Wenger DA and Sands MS (2005). "AAV2/5 vector expressing 645 galactocerebrosidase ameliorates CNS disease in the murine model of globoid-cell leukodystrophy more 646 efficiently than AAV2." Molecular Therapy 12(3): 422-430. 647 Lundgaard I, Osorio MJ, Kress BT, Sanggaard S and Nedergaard M (2014). "White matter astrocytes in health 648 and disease." Neuroscience 276: 161-173. 649 Martinez B and Peplow PV (2018). "Neuroprotection by immunomodulatory agents in animal models of 650 Parkinson's disease." Neural Regen Res 13(9): 1493-1506. 651 Miron VE, Jung CG, Kim HJ, Kennedy TE, Soliven B and Antel JP (2008). "FTY720 modulates human 652 oligodendrocyte progenitor process extension and survival." Ann Neurol 63(1): 61-71. 653 Miron VE, Kuhlmann T and Antel JP (2011). "Cells of the oligodendroglial lineage, myelination, and 654 remyelination." Biochim Biophys Acta 1812(2): 184-193. 655 Miron VE, Ludwin SK, Darlington PJ, Jarjour AA, Soliven B, Kennedy TE and Antel JP (2010). "Fingolimod 656 (FTY720) enhances remyelination following demyelination of organotypic cerebellar slices." Am J Pathol 657 176(6): 2682-2694. 658 Misslin C, Velasco-Estevez M, Albert M, O’Sullivan SA and Dev KK (2017). "Phospholipase A2 is involved in 659 galactosylsphingosine-induced astrocyte toxicity, neuronal damage and demyelination." PLOS ONE 12(11): 660 e0187217. 661 Mohri I, Taniike M, Taniguchi H, Kanekiyo T, Aritake K, Inui T, Fukumoto N, Eguchi N, Kushi A, Sasai H, Kanaoka 662 Y, Ozono K, Narumiya S, Suzuki K and Urade Y (2006). "Prostaglandin D2-mediated microglia/astrocyte 663 interaction enhances astrogliosis and demyelination in twitcher." J Neurosci 26(16): 4383-4393. 664 Molofsky AV, Krencik R, Ullian EM, Tsai H-h, Deneen B, Richardson WD, Barres BA and Rowitch DH (2012). 665 "Astrocytes and disease: a neurodevelopmental perspective." Genes & development 26(9): 891-907. 666 Münzel JE and Williams A (2013). "Promoting Remyelination in Multiple Sclerosis—Recent Advances." Drugs 667 73(18): 2017-2029. 668 Nicaise AM, Bongarzone ER and Crocker SJ (2016). "A microglial hypothesis of globoid cell leukodystrophy 669 pathology." J Neurosci Res 94(11): 1049-1061. 670 O'Sullivan C and Dev KK (2015). "Galactosylsphingosine (psychosine)-induced demyelination is attenuated by 671 sphingosine 1-phosphate signalling." J Cell Sci 128(21): 3878-3887. 672 O'Sullivan C, Schubart A, Mir AK and Dev KK (2016). "The dual S1PR1/S1PR5 drug BAF312 (Siponimod) 673 attenuates demyelination in organotypic slice cultures." J Neuroinflammation 13: 31. 674 Potter GB and Petryniak MA (2016). "Neuroimmune mechanisms in Krabbe's disease." Journal of Neuroscience 675 Research 94(11): 1341-1348. 676 Pritchard AJ, Mir AK and Dev KK (2014). "Fingolimod attenuates splenocyte-induced demyelination in 677 cerebellar slice cultures." PLoS One 9(6): e99444. 678 Rafi MA, Rao HZ, Luzi P, Curtis MT and Wenger DA (2012). "Extended normal life after AAVrh10-mediated gene 679 therapy in the mouse model of Krabbe disease." Mol Ther 20(11): 2031-2042. 680 Sheridan GK and Dev KK (2014). "Targeting S1P receptors in experimental autoimmune encephalomyelitis in 681 mice improves early deficits in locomotor activity and increases ultrasonic vocalisations." Sci Rep 4: 5051. 682 Slowik A, Schmidt T, Beyer C, Amor S, Clarner T and Kipp M (2015). "The sphingosine 1-phosphate receptor 683 agonist FTY720 is neuroprotective after cuprizone-induced CNS demyelination." Br J Pharmacol 172(1): 80- 684 92. 685 Smith B, Galbiati F, Castelvetri LC, Givogri MI, Lopez-Rosas A and Bongarzone ER (2011). "Peripheral 686 neuropathy in the Twitcher mouse involves the activation of axonal caspase 3." ASN Neuro 3(4). 687 Smith BR, Santos MB, Marshall MS, Cantuti-Castelvetri L, Lopez-Rosas A, Li G, van Breemen R, Claycomb KI, 688 Gallea JI, Celej MS, Crocker SJ, Givogri MI and Bongarzone ER (2014). "Neuronal inclusions of alpha- 689 synuclein contribute to the pathogenesis of Krabbe disease." J Pathol 232(5): 509-521. 690 Suzuki K (1998). "Twenty Five Years of the “Psychosine Hypothesis”: A Personal Perspective of its History and 691 Present Status." Neurochemical Research 23(3): 251-259. 692 Suzuki K (2003). "Globoid cell leukodystrophy (Krabbe's disease): update." J Child Neurol 18(9): 595-603. 693 Suzuki K and Suzuki Y (1970). "Globoid cell leucodystrophy (Krabbe's disease): deficiency of galactocerebroside 694 ɴ-galactosidase." Proceedings of the National Academy of Sciences of the United States of America 66(2): 695 302-309.
29 | Page
696 Wenger DA, Rafi MA and Luzi P (2016). "Krabbe disease: one hundred years from the bedside to the bench to 697 the bedside." Journal of Neuroscience Research 94: 982-989. 698 Wicks SE, Londot HE, Zhang B, Dowden J, Klopf-Eiermann J, Fisher-Perkins JM, Trygg CB, Scruggs BA, Zhang X, 699 Gimble BA, Bunnell BA and Pistell PJ (2011). "Effect of intrastriatal mesenchymal stromal cell injection on 700 progression of a murine model of krabbe disease." Behav Brain Res 225(2): 415-425. 701 Won JS, Singh AK and Singh I (2016). "Biochemical, cell biological, pathological, and therapeutic aspects of 702 Krabbe's disease." J Neurosci Res 94(11): 990-1006. 703 Yuan A, Rao MV, Veeranna and Nixon RA (2012). "Neurofilaments at a glance." Journal of Cell Science 125(14): 704 3257.
705
30 | Page