Dimethylsulfoxide Inhibits Oligodendrocyte Fate Choice of Adult Neural Stem and Progenitor Cells

Several clinical trials address demyelinating diseases via transplantation of mesenchymal stromal cells (MSCs). Published reports detail that administration of MSCs in patients may provide a beneficial immunomodulation, and that factors secreted by MSCs are potent inducers of oligodendrogenesis. Dimethylsulfoxide (DMSO) is widely used in life science and medicine as solvent, vehicle or cryoprotectant for cells used in transplantation. Importantly, most transplantation protocols do not include the removal of DMSO before injecting the cell suspension into patients. This indifferent application of DMSO is coming under increasing scrutiny following reports investigating its potential toxic side-effects. While the impact of DMSO on the central nervous system (CNS) has been partially studied, its effect on oligodendrocytes and oligodendrogenesis has not been addressed yet. Consequently, we evaluated the influence of DMSO on oligodendrogenesis, and on the pro-oligodendrogenic effect of MSCs’ secreted factors, using adult rat neural stem and progenitor cells (NSPCs). Here, we demonstrate that a concentration of 1% DMSO robustly suppressed oligodendrogenesis and drove the fate of differentiating NSPCs toward astrogenesis. Furthermore, the pro-oligodendrogenic effect of MSC-conditioned medium (MSCCM) was also nearly completely abolished by the presence of 1% DMSO. In this condition, inhibition of the Erk1/2 signal transduction pathway and high levels of Id2 expression, a specific inhibitor of oligodendrogenic differentiation, were detected. Furthermore, inflammatory demyelinating diseases may even potentiate the impact of DMSO on oligodendrogenesis. Our results demonstrate the imperative of considering the strong anti-oligodendrogenic activity of DMSO when designing future clinical trial protocols.


INTRODUCTION
In young adults, the most common neurological disease leading to permanent disability is multiple sclerosis (MS). The latter is a neuroinflammatory disorder of the central nervous system (CNS) characterized by the progressive destruction of myelin sheaths. Although MS patients can recover from their first demyelinating episodes through a process known as remyelination, the regenerative capacity of the CNS declines through the course of the disease. This leads to physical and cognitive deficits over time due to insufficient axon insulation and neuronal loss. Despite new drug developments, the repair of CNS damage still constitutes a challenging problem (Wingerchuk and Carter, 2014). As our understanding of the mechanisms underlying the process of remyelination grows, there is increasing interest in cell-based therapies to address the issue.
These properties encouraged the development of multiple clinical trials involving the application of MSCs to patients suffering from a demyelinating disease (summary of published studies of MSC transplantation in MS in Scolding et al., 2017). However, diverging results between studies raise questions about the efficacy of this MSC-based strategy and warrant standardized protocols, which should include the origin of cell product, the route of delivery, the dose, and the trial design (Scolding et al., 2017). Consideration should also be given to the presence or removal of the cryoprotectant dimethylsulfoxide (DMSO) from the cell preparation administered. While it has been demonstrated that a standard dose of 10% DMSO in the cell suspension administered is generally well tolerated, adverse reactions to cell applications could significantly be reduced when DMSO was removed or its concentration reduced prior to injection (Windrum et al., 2005). Moreover, case reports describing neurological adverse events like epileptic seizures, stroke, encephalopathy or leukoencephalopathy, following administration of DMSO suggest that it might be harmful for the CNS (Dhodapkar et al., 1994;Higman et al., 2000;Hoyt et al., 2000;Bauwens et al., 2005). Previous studies explored the influence of DMSO on neurons and astrocytes and reported concerning results about cell toxicity even at low DMSO concentrations (Hanslick et al., 2009;Tamagnini et al., 2014;Yuan et al., 2014;Zhang et al., 2017). However, DMSO's effect on other CNS resident precursor cells has not been addressed yet.
Our study demonstrates that a concentration of 1% DMSO almost completely obliterates the generation of oligodendrocytes from adult rat NSPCs in favor of astrocytic differentiation. These results substantiate the assumption that DMSO could have a negative impact on oligodendrogenesis, and we therefore call for further investigation of its potential hazard when applied in the context of demyelinating disorders.

Animal Works
Experiments were performed in accordance to the guidelines of the "Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes." According to the European Directive 2010/63/EU Article 3 and the Austrian legislation for experiments on living animals Tierversuchsgesetz 2012 §2c, no additional approval was required for the killing of animals with the aim of collecting tissues.

MSC Cultures and MSCCM
Mesenchymal stromal cell cultures and MSC-conditioned medium (MSCCM) were generated as described by Rivera et al. (2006). In brief, bone marrow plugs from tibias and femurs of 6-8 weeks old female Fischer-344 rats were harvested and mechanically dissociated in α-MEM. After centrifugation and resuspension, dissociated bone marrow cells were seeded at 1 × 10 6 cells/cm 2 in α-MEM containing 10% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin (thereafter aMEM) and grown in a humidified incubator at 37 • C with 5% CO 2 . Medium was changed every third day until a confluent cell layer was reached. For MSCCM production, MSCs were seeded at a density of 12,000 cells/cm 2 in aMEM. After 3 days of incubation, the supernatant (conditioned medium) was collected and filtered using a 0.22 µm pore filter (Rivera et al., 2006).

NSPC Cultures
Neural stem and progenitor cells cultures from the adult hippocampus (HC) and subventricular zone (SVZ) were established as described by Wachs et al. (2003). Briefly, 6-8 weeks old female Fischer-344 rats were decapitated and their HC and SVZ dissected. The tissue was transferred to 4 • C cold PBS containing 4.5 mg/L D-glucose, then mechanically minced and enzymatically digested. Dissociated cells were collected and resuspended in Neurobasal A medium (NBA, Gibco) supplemented with B27 (Gibco), 2 mM L-glutamine, 100 U/mL penicillin, 100 µg/mL streptomycin, as well as 2 µg/mL heparin, 20 ng/mL human recombinant EGF and 20 ng/mL human recombinant FGF (thereafter referred to as NBA+all). Neurosphere cultures were kept in a humidified incubator with 5% CO 2 at 37 • C and half of the medium was changed every third day.
For passaging, the medium containing floating neurospheres was centrifuged at 120 × g for 5 min. The supernatant was discarded and the pellet dissociated with Accutase (PAN Biotech). The single cells were then reseeded into NBA+all at a density of 1 × 10 5 cells/mL. Passaging took place every 7-9 days.
For freezing, the cell suspension was centrifuged 3 days after passaging and the supernatant was discarded. The cells were then transferred into cryomedium (10% DMSO, 20% FBS, 70% NBA supplemented with B27, 2 mM L-glutamine, 100 U/mL penicillin, 100 µg/mL streptomycin) and frozen at −150 • C until further use. After thawing, cells were immediately washed with NBA+all to remove remnants of the cryomedium, and cultured in NBA+all.
All NSPCs used in the experiments were frozen at passage 1 or 2, and employed for the various experiments in passage number 2-6. No difference in fate choice was observed during differentiation between fresh cells, and cultures derived from frozen stocks (data not shown). Data provided in the main figures were generated with hippocampal NSCs. Differentiation of NSCs obtained from the SVZ was not significantly different (Supplementary Figure S1).

Treatment of NSPCs
Neurospheres were dissociated using Accutase (PAN Biotech) and seeded onto 100 µg/mL poly-L-ornithine and 5 µg/mL laminin coated coverslips at a density of 8000-10,000 cells/cm 2 in aMEM. After 16 h, medium was replaced with aMEM or MSCCM containing the final concentration of DMSO. After 3 or 6 days of differentiation, the NSPCs were fixed with 4% paraformaldehyde and processed for immunofluorescence stainings.
For RNA-isolation, NSPCs were grown as mentioned above on poly-L-ornithine and laminin coated petri dishes in aMEM and MSCCM containing either 0 or 1% of DMSO. After 3 days of differentiation, NSPCs were lysed with Tri Reagent (Sigma-Aldrich) and total RNA was isolated using the RNeasy plus mini RNA isolation kit (Qiagen) (see below).
For protein isolation, NSPCs were seeded onto poly-Lornithine and laminin coated petri dishes for 16 h. After that, the NSPCs were stimulated with aMEM or MSCCM with 0 or 1% DMSO. After 10 min of treatment, protein isolation was performed using a RIPA-buffer.

Protein Extraction and Western Blots
Neural stem and progenitor cells were treated as mentioned above and after washing with PBS, cells were lysed with RIPA buffer supplemented with protease inhibitors (cOmplete, Roche) and phosphatase inhibitors (PhosSTOP, Roche). The suspensions were homogenized through a 27-gauge needle and samples placed on ice for 30 min. Protein concentrations of lysates were determined with a Pierce BCA Protein Assay Kit (Thermo Fisher Scientific). Equal amounts (10 µg) of protein samples were loaded on Mini-PROTEAN TGX Stain-free precast gel (Bio-Rad) and run with 120 V. Precision Plus Protein Dual Color Standard (Bio-Rad) as well as SeeBlue Plus2 Pre-stained Protein Standard (Thermo Fisher Scientific) were used as protein standard markers. Blotting was performed using a Trans-Blot Turbo Transfer System (Bio-Rad) with Trans-Blot Turbo Mini Transfer Packs (Bio-Rad).

Quantitative Gene Expression Analyses
Neural stem and progenitor cells were differentiated for 3 days as mentioned above. RNA-extractions were performed using TRI Reagent (Sigma) followed by the use of an RNeasy plus mini RNA isolation kit (Qiagen). The cDNA synthesis was performed with iScript Reverse Transcription Supermix for RT-qPCR (Bio-Rad). Quantitative gene expression analyses were performed using TaqMan RT-PCR technology. Technical duplicates containing 10 ng of cDNA were amplified with the GoTAQ Probe qPCR Master Mix (Promega) using a two-step cycling protocol (95 • C for 15 s, 60 • C for 60 s; 40 cycles, Bio-Rad CFX 96 Cycler).

Statistical Analysis
Data are presented as means ± SD. Statistical analysis was performed using RStudio (RStudio Inc., Boston, MA, United States). Data sets were analyzed according to the general factorial design included in the GFD-package in R (Friedrich et al., 2017) using the ANOVA-type approximation for the calculation of p-values. P-values of <0.05 were considered significant. Significance: ns, p ≥ 0.05; * p < 0.05; * * p < 0.01; * * * p < 0.001.
To exclude the possibility that the loss of oligodendrocytes results from an early selection through lineage-specific apoptosis, we quantified the frequency of active Caspase 3 in NG2+, Olig2+  Table S1). These results suggest that DMSO blocked the oligodendrogenic activity of MSCCM by suppressing the intrinsic capacity of NSPCs to undergo oligodendroglial differentiation in favor of an astrocytic fate.
In addition to their absolute levels of expression, the ratio between Olig2 and Id2 in the NSPCs is also crucial for fate decision. We have previously reported that pretreatment with MSCCM during proliferation enhanced the Olig2/Id2 ratio in NSPCs and this high ratio promotes an oligodendrocytic fate (Steffenhagen et al., 2012). Comparing the ratio calculated for various conditions used to differentiate the NSPCs, it is conspicuous that the addition of 1% DMSO, either in aMEM or in MSCCM, robustly decreased the Olig2/Id2 ratio, in correlation with the inhibition of the oligodendrogenic activity ( Figure 3B).
of a well-known oligodendrogenic pathway by the addition of DMSO (Figure 4).

DISCUSSION
For decades, DMSO has probably been the most utilized solvent for lipophilic agents in in vitro and in vivo studies. Due to its unique chemical properties, DMSO enables miscibility of hydrophobic substances and serves as a drug carrier enabling penetration into mammalian tissues. In addition, DMSO is widely used as cryoprotectant, e.g., for storage of stem cells. Unlike previous assumptions describing DMSO as inert in low concentrations, recent reports suggested an impact on the cell cycle, differentiation patterns, cell survival and epigenetic changes, even for concentration of DMSO below 1% (Santos et al., 2003;Iwatani et al., 2006). In the past years, additional studies have demonstrated that DMSO induces neurotoxic effects (Hanslick et al., 2009;Yuan et al., 2014), neurophysiological alterations in pyramidal neurons (Tamagnini et al., 2014) and activation of astrocytes (Zhang et al., 2017). Even though concerns over potential risks of using DMSO are growing, its impact on NSPCs residing in the CNS has not been carefully addressed yet.
In the present study, we investigated the impact of DMSO on the differentiation and glial fate decision of NSPCs. As previously reported (Rivera et al., 2006), under standard culture conditions (aMEM), the vast majority of NSPCs spontaneously differentiated into the astrocytic lineage, whereas addition of MSCCM strongly directed the fate choice toward oligodendrocytes. We could demonstrate that this pro-oligodendrogenic effect was inhibited to a large extent by the presence of DMSO during differentiation. This inhibitory activity of DMSO was not limited to NSPCs from the hippocampus, but also applied to NSPCs from the other adult neurogenic niche, i.e., the SVZ. We further observed a progressive inhibition of oligodendrogenesis with increasing concentrations of DMSO, however, significant inhibition in our in vitro conditions was observed at a concentration of 1% DMSO. As the detailed mode of action of DMSO on cell fate remains elusive, a significant inhibition of oligodendrogenesis at lower concentrations cannot be ruled out in the presence of additional factors or under pathological conditions. We further investigated key determinants of oligodendrogenesis and found that the application of DMSO elevated the expression of Id2 in differentiating NSPCs cultivated either in aMEM, or in MSCCM. An overall decrease in the Olig2/Id2 ratio resulted in a shift of NSPCs fate decision away from oligodendrocytes and toward astrocytes. We also investigated the extracellular signal regulated kinases 1/2 (Erk1/2) pathway, which has been shown to regulate oligodendrogenic differentiation in NSPCs (Hu et al., 2004;Shi et al., 2014). We examined the phosphorylation status of Erk1/2 and detected a strong upregulation of the phosphorylated isoform following application of MSCCM. In contrast, the addition of DMSO in MSCCM blocked the phosphorylation of Erk1/2. Whether Erk1/2 dephosphorylation directly modulates Id2 expression under the influence of DMSO still needs to be evaluated.
Although remyelination is mainly achieved by OPC differentiation (Franklin and Goldman, 2015), NSPCs in the adult neurogenic niches have been shown to proliferate and migrate toward lesion sites and locally differentiate into oligodendrocytes as well (Menn et al., 2006;Etxeberria et al., 2010;Jablonska et al., 2010;Kazanis et al., 2017) As our results demonstrate strong anti-oligodendrogenic properties mediated by DMSO for cells of different neurogenic niches, even in the presence of MSCCM, the simultaneous application of MSCs and DMSO in transplantations might inhibit the MSC-mediated pro-oligodendrogenic activity. Furthermore, it would be of interest to evaluate, if OPC differentiation is similarly affected by DMSO, as documented in NSPCs, since upregulation of Id2 and dephosphorylation of Erk1/2 have been shown to block differentiation of OPCs as well (Wang et al., 2001;Fyffe-Maricich et al., 2011;Guardiola-Diaz et al., 2012).
In summary, our study demonstrates a robust antioligodendrogenic effect of DMSO on fate decision when present during NSPCs' differentiation in vitro. Due to its broad use in cell culture, in vivo experiments and clinical trials, future studies evaluating the differentiation of NSPCs, oligodendrogenesis or remyelination should consider the impact of DMSO carefully and, whenever its removal is not possible, emphasize the use of appropriate controls.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
Ethical review and approval was not required for the animal study because the legislation does not require ethics approval for the killing of animals with the aim of collecting tissues.

AUTHOR CONTRIBUTIONS
AO'S, SL, PR, and LB conducted and analyzed the experiments. AO'S, PR, and LB performed the statistical analysis. AO'S, FR, and SC-D wrote the manuscript. All authors contributed to the conception, design, and finalizing of the study, and approved the submitted version.

ACKNOWLEDGMENTS
The authors are grateful for the support of Georg Zimmermann (Department of Neurology, Spinal Cord Injury and Tissue Regeneration Center Salzburg, Paracelsus Medical University, Salzburg, Austria) for his support to the statistical analysis.
FIGURE S2 | DMSO and MSCCM decrease DCX expression. RT-PCRs of DCX gene expression were performed on NSPCs differentiating for 3 days. Relative mRNA expression level of DCX detected in NSPCs cultivated in aMEM or MSCCM, with or without 1% DMSO respectively. Data are shown as mean ± SD. Asterisks mark significant difference compared to aMEM, * p < 0.05, n.s.: not significant. TABLE S1 | DMSO does not induce specific cell death in cells of oligodendrocyte lineage. Percentages of NG2/Casp3 + , Olig2/Casp3 + and NG2/Olig2/Casp3 + cells in relation to their respective population (NG2 + , Olig2 + , NG2/Olig2 + ) after 3 days of differentiation. Percentages of CNP/Casp3+ and GFAP/Casp3+ cells in relation to their respective population (CNP+ and GFAP+). P-values comparing fractions of Casp3 + cells of specific NG2 + , Olig2 + and NG2/Olig2 + cell populations in aMEM or MSCCM with and without 1% DMSO. Data represented as mean ± SD. n = 3. ND: not determined. DMSO did not significantly affected Casp3 frequency in the subpopulations.