In Vitro Characterization of Motor Neurons and Purkinje Cells Differentiated from Induced Pluripotent Stem Cells Generated from Patients with Autosomal Recessive Spastic Ataxia of Charlevoix-Saguenay

Autosomal recessive spastic ataxia of Charlevoix-Saguenay (ARSACS) is an early-onset neurodegenerative disease mainly characterized by spasticity in the lower limbs and poor muscle control. The disease is caused by mutations in the SACS gene leading in most cases to a loss of function of the sacsin protein, which is highly expressed in motor neurons and Purkinje cells. To investigate the impact of the mutated sacsin protein in these cells in vitro, induced pluripotent stem cell- (iPSC-) derived motor neurons and iPSC-derived Purkinje cells were generated from three ARSACS patients. Both types of iPSC-derived neurons expressed the characteristic neuronal markers β3-tubulin, neurofilaments M and H, as well as specific markers like Islet-1 for motor neurons, and parvalbumin or calbindin for Purkinje cells. Compared to controls, iPSC-derived mutated SACS neurons expressed lower amounts of sacsin. In addition, characteristic neurofilament aggregates were detected along the neurites of both iPSC-derived neurons. These results indicate that it is possible to recapitulate in vitro, at least in part, the ARSACS pathological signature in vitro using patient-derived motor neurons and Purkinje cells differentiated from iPSCs. Such an in vitro personalized model of the disease could be useful for the screening of new drugs for the treatment of ARSACS.


Introduction
Autosomal recessive spastic ataxia of Charlevoix-Saguenay (ARSACS) is a childhood-onset hereditary neurodegenerative disease. ARSACS is considered the second most frequent ataxia after the Friedreich ataxia [1]. Described for the first time in the 1970s [2] with a carrier prevalence of 1/22 in the Charlevoix-Saguenay region of the Quebec Province, ARSACS is now found in more than twenty countries worldwide [3,4]. Different causative mutations in the SACS gene are linked to the disease. To date, more than 200 mutations have been identified in this gene [5]. The SACS gene encodes the protein sacsin, highly expressed in Purkinje cells and motor neurons (MNs), which is involved in molecular chaperoning, mitochondrial transport and integrity, and neurofilament assembly [6]. Sacsin knockout mice recapitulated at least in part histopathological and neurological features of ARSACS, pointing out to a sacsin loss-of-function disease mechanism [7]. ARSACS is characterized by cerebellar, pyramidal, and neuropathic involvement and remains incurable [8]. Noteworthy, research on ARSACS patient fibroblasts and mice with decreased SACS gene expression showed mitochondrial abnormalities, with hyperfused mitochondria and a reduction in Drp1-mediated mitochondrial fission [9][10][11]. Mitochondria autophagy was studied on a commercial neuron-like cell line, and an alteration of this mechanism was discovered [12]. Studies conducted using ARSACS mouse models also indicated an alteration in the organization of the intermediate filament network with an accumulation of neurofilaments [7,13]. Although the actual ARSACS animal models are still valuable to investigate the role of sacsin in the pathology of the disease, more reliable humanderived cellular models need to be generated to better understand the underlying pathophysiology of ARSACS in specific cell types. To overcome the lack of human tissue accessibility, interest has now focused on the differentiation of patientderived induced pluripotent stem cells (iPSCs) into cerebellar cells such as Purkinje cells or granule cells [14][15][16]. A cerebellum organoid model has also been developed to recapitulate the complexity of cerebellar tissue [17]. However, organoids sometimes cause some reproducibility issues [18][19][20].
In this paper, we propose reproducible methods to differentiate ARSACS patient-derived iPSCs into MNs and Purkinje cells. We also show for the first time the detection of ARSACS pathological features in patient-derived neuronal cells to confirm the pathological effects caused by the observed impairment of sacsin in mouse models or patient-derived nonneuronal cells.

Materials and Methods
2.1. Cell Culture. This study was approved by the ethical committee of the CHU de Québec-Université Laval (#2019-4444). The human cells used in these experiments were obtained following informed consent from the donors.
As previously described [21], healthy human dermal fibroblasts were isolated from skin biopsies after breast reductive surgery and cultured in DMEM-F12 medium supplemented with 10% fetal calf serum (FCS; Wisent, St-Bruno, Canada) and 25 μg/ml of penicillin/gentamycin (MilliporeSigma, Oakville, Canada). Human epineural fibroblasts were obtained after peripheral nerve surgery from a small piece of nerve epineurium and cultured in DMEM-F12 supplemented with 20% FCS and antibiotics (100 U/ml penicillin and 25 μg/ml gentamycin).
The purity of MNs was assessed on 3 control iPS cell lines at day 12 of differentiation. To quantify the number of islet1 and β3Tubulin-positive cells, 50,000 cells/cm 2 were seeded in 24-well plates, and 6 wells were counted by immunocytochemistry, with 3 counts per well on randomly selected area (between 75 and 200 cells/area) using an Imager M2 stereo investigator (Carl Zeiss Microscopy). The results were expressed as the percentage of islet1 and β3Tubulin-positive cells in each of the 3 different cell lines normalized against iPSCs that do not express both markers.

Differentiation of iPSCs into Purkinje Cells.
Differentiation of iPSCs into Purkinje cells was conducted following modifications of protocols established from [15,16,[24][25][26][27][28][29][30][31][32][33][34]. The iPSCs were dissociated with StemPro Accutase (Thermo Fisher Scientific), combined with 5 μM Y-27632 (Abcam), and diluted in DMEM-F12 medium added with 5 μM Y-27632 without serum. The iPSCs were plated at a density of 90,000 cells/cm 2  were added to the medium, but at day 4, SB431542 and LDN19318 were removed and CHIR99021 was removed at day 7. From day 2 to day 9 of differentiation, the medium was supplemented with 7 μg/ml insulin and 35 ng/ml βFGF. At days 7 and 8, 10 μM cyclopamine was added to the medium. From days 9 to 18 of differentiation, 10 μM DAPT was added to the medium. At day 11, cells were dissociated with StemPro Accutase (Thermo Fisher Scientific), combined with 5 μM Y-27632 (Abcam), diluted in DMEM-F12 medium without serum and containing 5 μM Y-27632, and seeded at 400,000 cell/cm 2 on 10 μg/ml poly-D-lysine (MilliporeSigma) plates for cell characterization or 285,000 cells/cm 2 over the 3D sponge model. At day 11, medium was added with 100 ng/ml NT-3 (FroggaBio) until day 19, 10 pg/ml GDNF until the end, and 1 nM urocortin until day 15 (R&D System, Oakville, Canada). From days 12 to 21 of differentiation, medium was supplemented with 30 nM T3 (3,3 ′ ,5-triiodo-Lthyronine sodium salt, MilliporeSigma). From days 13 to 18, 50 ng/ml FAIM2/LFG protein (Abcam) was added to the medium. From days 15 to 25 of differentiation, 1 nM corticotropin-releasing factor (CRF, Bio-Techne, Toronto, Canada) was added daily to the medium for a 12-hour period. From day 19 until the cell peeled off in the 2D experiment or to day 53 in the 3D experiment, the medium was supplemented with 20 ng/ml BDNF. The purity of Purkinje cells was assessed on 3 control iPS cell lines at day 14 of differentiation. To quantify the number of KIRREL2 and β3Tubulin-positive cells, 50,000 cells/cm 2 were seeded in 24-well plates, and 6 wells were counted by immunocytochemistry, with 3 counts per well on randomly selected area (between 75 and 200 cells/area) using an Imager M2 stereo investigator (Carl Zeiss Microscopy). The results were expressed as the percentage of KIRREL2 and β3Tubulin-positive cells in each of the 3 different cell lines, normalized against iPSCs that do not express both markers.

Preparation of 3D Cell Culture Substrates.
Collagen sponges were prepared as previously described [37,38] but without chondroitin 4-6 sulphate. Briefly, type I and III bovine collagen (Symatese, Chaponost, France) and chitosan (Kemestrie, Sherbrooke, Canada) were dissolved in 0.1% acetic acid and mixed, and 0.5 ml of the mixture was poured into 12-well plates and frozen at -80°C for 1 h. The frozen plates were then lyophilized in a vacuum lyophilizer (Dura-Stop Microprocessor Controlled Tray Freeze-Dryer; FTS Systems, Stone Ridge, NY).

MN Differentiation.
To culture the MNs in 3D, 210,000 dermal human fibroblasts/cm 2 were seeded onto the collagen/chitosan sponge and cultured for 1 week.

Stem Cells International
The sponge was flipped upside down and put at the airliquid interface, and 100,000 human epineural fibroblasts/ cm 2 were seeded onto the sponge and cultured for one week with DMEM-F12 supplemented with 10% FCS and antibiotics (100 U/ml penicillin and 25 μg/ml gentamycin). Sponge was rinsed, and 260,000 MNs/cm 2 at day 11 of differentiation were seeded on the dermal fibroblast side. On day 13 of differentiation, the culture medium was composed of DMEM-F12 : neurobasal A medium (vol : vol), 1% N2 supplement, 1% B-27 supplement minus vitamin A, 50 μg/ml ascorbic acid, 5 μM Y-27632, 1% L-alanyl-Lglutamine, 1% MEM nonessential amino acids, 0.1% trace element A, 0.1% trace element B, 0.1% trace element C, 10 μM DAPT, 10 ng/mL GDNF, 20 ng/mL BDNF, 10 ng/ ml NT-3, and antibiotics (100 U/ml penicillin and 25 μg/ ml gentamycin). One week after the MN seeding, ARSACS or normal iPSC-derived Schwann cells were seeded on the same side as the epineural fibroblasts; DAPT was removed, and 10 ng/ml NGF was added until the end of the culture 8 weeks later (for a total of 74 days of iPS-derived MN differentiation).
2.6. Immunofluorescence. Cells cultured on plastic or on 3D substrates were fixed, respectively, 20 min and 1 h at 4°C with 4% paraformaldehyde and washed with cold PBS containing 0.3% Triton X-100 (Bio-Rad, Mississauga, Canada). 3D substrates were cut into 15 mm 2 pieces or into 30 μm cross-sections. Then, cells and 3D substrates were incubated 20 min at room temperature in an immunofluorescence staining buffer containing cold PBS, 0.3% Triton X-100 (Bio-Rad), and 5% serum. Cells and 3D substrates were incubated overnight at 4°C in the immunofluorescence staining buffer containing selected primary antibodies and were washed three times with PBS containing 0.3% Triton X-100. Cells and 3D substrate cross-sections were then incubated 1 h at room temperature in the immunofluorescence staining buffer containing selected secondary antibodies diluted (1 : 500). Finally, cells and 3D substrate cross-sections were washed with PBS -0.3% Triton X-100 and one last time with PBS before mounting with Fluoromount-G containing DAPI (Electron Microscopy Sciences, Hatfield, PA). Imaging was carried out using a LSI 700 confocal microscope with Zeiss Axio Imager (Carl Zeiss Microscopy, Jena, Germany).
The primary antibodies used were mouse anti-ß3Tub

Western Blot Analysis.
To extract total proteins from Purkinje cells cultured on 3D substrates, samples were lysed in 1x RIPA buffer (Abcam) and 1x Pierce protease and phosphatase inhibitor mini tablet (Thermo Fisher Scientific). The protein amount extracted from each Purkinje cell substrate was assessed with a Bio-Rad Protein Assay (Bio-Rad Laboratories Inc., Hercules, CA). An equal amount of lysates was loaded on a stacking gel and run in 8% or 10% size-fractioned SDS-polyacrylamide gel electrophoresis. Proteins were transferred to a polyvinylidene difluoride blotting membrane (Bio-Rad). Membranes were blocked in 5% nonfat dried milk containing 0.05% Tween 20 (VWR, Radnor, PA) and Tris-buffered saline (TBS) for 1 hour. Membranes were then probed overnight at 4°C with primary antibodies such as rabbit anti-sacsin (1 : 1000; Abcam), mouse anti-DRP1 (1 : 1000; Abcam), rabbit anti-vinculin (1 : 1000; Cell Signaling Technology, Whitby, Canada), and mouse anti-β-actin (1 : 1000; Abcam) diluted in the blocking solution. Membranes were washed in TBS-0.05% Tween 20 and incubated for 1 hour at room temperature with secondary antibodies such as goat anti-rabbit IgG-HRP (1 : 5000; Jackson ImmunoResearch Inc.) and goat anti-mouse IgG-HRP (1 : 5000; Jackson ImmunoResearch Inc.) diluted in blocking solution. Membranes were incubated with Amersham ECL Western Blotting Detection Reagent (GE Healthcare, Mississauga, Canada) to detect protein expression and were imaged with a Fusion Fx7 imager (Vilber Lourmat Sté, Collégien, France). The densitometric analyses of the bands were performed using ImageJ software (National Institutes of Health, Bethesda, MD). DRP1 protein expression was normalized to β-actin (42 kDa), but sacsin protein expression was normalized to vinculin (117 kDa) because of its high molecular weight (520 kDa).

Statistical
Analysis. Statistical analyses were performed using GraphPad Prism 9.0 software (GraphPad Software Inc., La Jolla, CA). Data are presented as mean ± standard error of the mean (SEM). Nonparametric one-way ANOVA comparison was used to compare the control group with ARSACS group for each type of neurons. Tests having a p value above 0.05 were determined statistically different.

Differentiation of iPSCs from ARSACS Patients into
MNs. iPS cells derived from 3 ARSACS patients and 3 healthy individuals were differentiated into MNs during 13 days for 2D culture in 12-well plates coated with poly-Dlysine. MN maturity and identity was assessed by immunofluorescence staining with general neuronal cell markers such as neurofilament M (NFM; 160 kDa) and β3-tubulin and markers specific to MNs through Islet1 and ChAT. Both ARSACS and control iPSC-derived MNs expressed all these markers, demonstrating the efficacy of our differentiation protocol (Figures 1(a)-1(d)). No difference was observed in the expression of these markers between MNs differentiated from ARSACS iPSCs compared to control iPSCs in 2D culture. The efficiency of iPSC differentiation into MN was 4 Stem Cells International assessed on control iPSC lines by counting the proportion of cells coexpressing Islet1 and β3-tubulin (Figure 1(m)). The C3 cell line gave a better purity of 83% ± 3:6, compared to the C2 cell line (60% ± 22:7) and C1 (23% ± 11). When cultured in a 3D environment for 74 days in total, MNs coexpressed NFM and ChAT (Figure 2(e)) as well as Islet1 with NFM (Figure 2(f)) as observed in 2D culture.  5 Stem Cells International using β3-tubulin or NFM, as well as specific Purkinje cell markers like GRID2, KIRREL2, parvalbumin, or calbindin (Figures 1(e)-1(l)). These markers were mostly expressed in the cell body, except for GRID2 that was also detected along some dendrites. Moreover, these cells did not express PCP2 at this early differentiation stage (Figures 1(k) and 1(l)). No difference was observed in the expression of these markers between Purkinje cells differentiated from ARSACS and control iPSCs in 2D culture.

Differentiation of iPSCs from ARSACS Patients into
The purity of Purkinje cells following iPSC differentiation was assessed by immunofluorescence analysis through counting the proportion of KIRREL2 and β3-tubulin coexpression in the cells. After 14 days of differentiation, more than 92% of iPSCs expressed both markers (Figure 1(n)).
A major limitation of growing Purkinje cells on plastic was their spontaneous detachment between day 15 and day 20 of in vitro differentiation, before reaching full maturation. To solve this problem, Purkinje cells were differentiated 11 6 Stem Cells International days on plastic (before the neurites were formed) and transferred on a 3D substrate to complete the needed differentiation period for an additional 6 weeks. Immunofluorescence analysis showed that differentiated Purkinje cells expressed the neuronal marker NFM (Figures 2(c) and 2(d)) as well as known Purkinje cell markers calbindin (Figure 2(a)), KIR-REL2 (Figure 2(b)), parvalbumin (Figure 2(c)), and PCP2 (Figure 2(d)). Moreover, some of these cells displayed a more stellar morphology with multiple dendrites (Figures 2(a) and 2(c)). Of note, these specific Purkinje cell markers were found to be homogeneously distributed and mostly detected along dendrites in long-term 3D cultures, whereas mainly detected in the cell body in 2D culture after 14 days of differentiation. In addition, PCP2 was not expressed at an earlier stage of differentiation in the 2D culture but was detected in the 3D model after 53 days of differentiation (Figures 1(k) and 1(l) vs. Figure 2(d)).

ARSACS Purkinje Cells and MNs Showed Decreased
Levels of Sacsin but Not of DRP1. The expression of DRP1 and sacsin proteins expressed by ARSACS and control Purkinje cells differentiated from iPSCs, cultured for 53 days in vitro on the 3D substrate, was assessed by immunofluorescence ( Figure 3). Both DRP1 and sacsin expressions seemed less intense in the ARSACS-derived Purkinje cells (Figures 3(a)

-3(d) versus Figures 3(e)-3(h)
). Semiquantitative measurements of DRP1 and sacsin expressions were then performed by Western blot analysis using whole protein extracts collected from ARSACS and control Purkinje cells differentiated from iPSCs ( Figure 4 and Figure S1 in Supplementary materials). A significant decrease of sacsin expression was detected in ARSACS iPSC-derived Purkinje cells (N = 3, p < 0:0005). This was measured both in cells differentiated 14 days in 2D culture as well as for cells differentiated in 3D collagen sponges for a total of 53 days. Note that healthy fibroblast expression of sacsin and DRP1 was found to be very low when compared to iPSC-derived Purkinje cells, which may therefore not interfere with the measurement of the total expression of these proteins in 3D experiments where both cell types are cocultured in the sponges.
Conversely, the expression of DRP1 was found to be similar in control versus Purkinje cells differentiated from ARSACS-derived iPSCs independent of the 2D or 3D culture methods. Similar results, i.e., a reduced sacsin expression level (N = 3, p < 0:0005) but comparable expression of DRP1, were observed in MNs differentiated from ARSACS-derived iPSCs cultured in 2D for 13 days as well as in 3D for 74 days. In this last condition, MNs were cocultured with healthy fibroblasts but with ARSACS iPSCderived Schwann cells in the ARSACS model (to promote axonal migration as previously described [39]).

Discussion
We have developed in this study simplified and easily reproducible protocols for the generation of MNs and

Stem Cells International
Purkinje cells differentiated from ARSACS patient-derived iPSCs, which successfully recapitulated the early pathological signature of the disease. Extracting living neurons from the cerebellum and other CSN tissues remains very difficult to achieve from postmortem human pathological tissue [4]. To overcome this drawback, several protocols have been developed to differentiate iPSCs into Purkinje cells to generate cerebellum organoids [41][42][43]. However, in the vast majority of these protocols, the coculture of Purkinje cells with animal cerebellar cells or on organotypic cerebellar slices are needed to facilitate the iPSC differentiation [14,15,17].
Here, the goal of this study was to develop a culture system entirely made of human cells, using Purkinje cells or MNs differentiated from iPSCs obtained from ARSACS patients.
Although it was previously showed that it was possible to generate an iPSC line from ARSACS patient dermal fibro-blasts, these patient-derived iPSCs were not differentiated into specific neuronal cells [44,45].
We showed for the first time that it was possible to differentiate ARSACS-iPSCs into MNs and Purkinje cells using the 3D culture system we developed, as well as to detect some characteristic ARSACS pathological features such as abnormal accumulation of NFM along the neurites. Using collagen sponges populated with dermal fibroblasts as a 3D substrate, which is known to promote long-term survival and neurite elongation of MNs [39] and to reproduce pathological features observed in other brain diseases such as neurofibromatosis [46] and amyotrophic lateral sclerosis [47,48], it was possible to enable long term (>53 days) of Purkinje cells and to promote abundant elaborated dendritic ramifications. To better mimic ARSACS cellular microenvironment around MNs, iPSC-derived Schwann cells were also cocultured with iPSC-MNs within the 3D substrate. Such condition allowing the coculture of Schwann cells with iPS-MN 3D   Control   C1  C2  C3  A1  A2  A3  C1  C2  C3  A1  A2    Purkinje cells differentiated from 3 healthy control iPSCs (C1, C2, C3) and 3 ARSACS iPSCs (A1, A2, A3) cultured in 2D and 3D, as well as on fibroblasts. DRP1 and sacsin were, respectively, normalized against actin and vinculin to control for equal loading quantification. A statistical analysis was performed using 1-way ANOVA nonparametric test, * p < 0:0005 (N = 3).

8
Stem Cells International iPSC-derived MNs was previously shown to be essential to enhance axonal migration and to promote myelin sheath formation around neurites [39,48]. We also showed that sacsin expression was greatly reduced in both iPSC-MNs and Purkinje cells derived from ARSACS patients, similarly to what it is already described in the literature in patients, ARSACS mice [9,11], and other in vivo studies [7]. We did not observe however misexpression of DRP1 in our model, previously shown to be slightly downregulated following loss of sacsin expression.
Finally, we showed for the first time NFM aggregate formation in ARSACS patient-derived Purkinje cell and MN neurites, as several studies confirmed when the sacsin protein is altered in mouse or human nonneuronal cells [7,13,40,49]. However, this effect was only detected after long-term maturation of these cells in our 3D culture conditions. Moreover, we did not observed accumulation of neurofilaments as seen in mouse Sacs -/motor neurons and human ARSACS brain [7]. This may be due to a differentiation period of MN and Purkinje cells that was too short in vitro to allow this characteristic to appear. This in vitro model, shown to recapitulate important ARSACS pathological features, could serve as a tool to better understand the disease mechanism and to test new therapeutic molecules or new strategies against ARSACS.

Conclusion
We present a new in vitro culture system allowing the differentiation of ARSACS patient-derived iPSCs into MNs and Purkinje cells and long-term coculture of these two neuronal types. In future studies, it will be interesting to complexify the model by incorporating other subpopulations of cerebellar cells such as granule cells, oligodendrocytes, or Bergmann glial cells.

Data Availability
Source data are provided in this paper. All relevant data are available from the authors upon reasonable request.

Disclosure
This manuscript was presented in the PhD thesis of Dr. Aurélie Louit at the Laval University in December 2021.

Conflicts of Interest
The authors have no conflict of interest to declare.