Oligodendrocyte development and CNS myelination are unaffected in a mouse model of severe spinal muscular atrophy

Abstract

The childhood neurodegenerative disease spinal muscular atrophy (SMA) is caused by loss-of-function mutations or deletions in the Survival Motor Neuron 1 (SMN1) gene resulting in insufficient levels of survival motor neuron (SMN) protein. Classically considered a motor neuron disease, increasing evidence now supports SMA as a multi-system disorder with phenotypes discovered in cortical neuron, astrocyte, and Schwann cell function within the nervous system. In this study, we sought to determine whether Smn was critical for oligodendrocyte (OL) development and central nervous system myelination. A mouse model of severe SMA was used to assess OL growth, migration, differentiation and myelination. All aspects of OL development and function studied were unaffected by Smn depletion. The tremendous impact of Smn depletion on a wide variety of other cell types renders the OL response unique. Further investigation of the OLs derived from SMA models may reveal disease modifiers or a compensatory mechanism allowing these cells to flourish despite the reduced levels of this multifunctional protein.

Introduction

Spinal muscular atrophy (SMA) is an often-fatal childhood-onset neurodegenerative disease of genetic origin (1). It is relatively common among paediatric genetic diseases, with a carrier frequency of 1 in 40 and an incidence of approximately 1 in 6,000–10,000 (2). The major pathological feature of SMA is the loss of alpha motor neurons in the spinal cord, translating to progressive proximal muscle weakening and atrophy (3,4).

Loss-of-function mutations or deletions in the Survival Motor Neuron 1 gene (SMN1) are the underlying cause of SMA (3). Humans also possess SMN2, an inverted duplication of SMN1 that harbours a C to T transition at position 6 within exon 7 that favours the exclusion of this exon from the transcript. This transcript yields a highly unstable, truncated protein (5). However, SMN2 also produces low amounts of full-length SMN, such that the copy number of SMN2 inversely correlates with disease severity (5,6).

Mice possess only one Smn gene and its loss leads to early embryonic lethality (7). A mouse model that closely recapitulates the most severe form of SMA was developed by Monani and colleagues, through supplementation of Smn knockout mice with a human SMN2 transgene (Smn^-/-;SMN2) (8). Study of this mouse model of severe SMA has facilitated the understanding of the role of Smn in different cell types, and its importance in the maintenance of lower motor neuron function. Insufficient Smn levels result in disorganization of the neuronal growth cone structure and subsequent stunting of axon extension (9,10). Due to the similarities in growth cone morphology between neurons and oligodendrocytes (OLs) (11), it is important to address the effect of low Smn levels in OLs. Additionally, Smn-depletion contributes to an increase in astrocyte reactivity, as well as functional defects in astrocyte signalling and growth factor secretion (12,13). As glial cell types within the CNS, the parallels between astrocytes and OLs are numerous, further justifying the need to address the role of Smn in OL function.

OLs are the myelinating cell type of the central nervous system (CNS). During development, oligodendrocyte precursor cells (OPCs) differentiate into mature OLs, which extend elaborate networks of processes that are ultimately responsible for the ensheathment of multiple axons (14). Functionally, OLs and the resultant myelin sheath are responsible for promoting the efficiency of axonal signal transduction, as well as serving to protect the surrounded axon from damage and providing metabolic support. Notably, myelin defects in the peripheral nervous system have been reported in multiple models of SMA, characterized by intrinsic defects in the Schwann cell, which is responsible for peripheral myelination (15,16). The key role played by the OL in supporting motor neuron function, along with the similarities to other affected cell types mentioned above, led us to investigate the effect of Smn depletion on OLs.

To determine the dependence of OLs on the Smn protein, we utilized the Smn^-/-;SMN2 (severe) mouse model. Investigation of OL differentiation and myelination revealed no difference in morphology or function in Smn depleted OLs compared to wild type controls, both in vitro and in vivo. Our data suggest that despite the multi-functionality and ubiquitous expression of the Smn protein, it does not play a key role in myelination of the CNS, at least in the context of SMA disease pathogenesis.

Results

Low Smn levels do not affect OPC growth or molecular differentiation in vitro

The ubiquitously expressed Smn protein is known to play a role in various cellular functions including the proper formation of elaborate axonal growth cones in motor neurons (9). The presence of dynamic, ‘growth cone-like’ process tips in OLs is increasingly recognized as the mechanism for orchestrating the highly branched network in the differentiating cell (11,14). Due to the commonalities between neuronal and OL growing tips, it is reasonable to speculate that Smn plays a role in the OL differentiation gamut. To investigate the role of Smn in this process, primary OLs harvested from Smn^-/-;SMN2 mice were compared to that of control littermates. Western blots confirmed that the amount of Smn protein was reduced in our mixed glial culture system (Supplementary Material, Fig. S1).

To assess OPC maturation, primary OLs were allowed to differentiate for three and six days in vitro (DIV3 and DIV6; Fig. 1). Progression through the OL differentiation paradigm can be tracked via the expression of many common maturation markers: NG2 proteoglycan for early, myelin-associated glycoprotein (MAG) for mid and myelin basic protein (MBP) for late stage differentiation. The proportion of OLs expressing each marker at the chosen time points in vitro was consistent between Smn^-/-;SMN2 and wild type cultures (Fig. 1B and C). Of note, the onset of expression of such differentiation markers occurred as expected, whereby NG2 levels decreased from DIV3 to DIV6 with a concomitant increase in MAG and MBP levels. It should be noted that OL differentiation followed a typical expression pattern such that NG2 is expressed early, and was lost as MAG was induced but with overlapping expression during this time. MBP followed MAG in expression; however MAG continued to be expressed throughout the remainder of OL differentiation. Therefore, the differentiation pattern of the OLs assessed from both wild type and Smn^-/-;SMN2 cultures was as expected, with almost complete co-expression of MAG and MBP by DIV6. In addition to the proportion of cells expressing each marker, it was also of interest to assess the relative amount of these vital myelin proteins within the positive-expressing OLs. The intensity of the fluorescence signal for MBP-positive OLs at DIV3 was quantified using the integrated density plugin of ImageJ software and no difference between genotypes was observed (Fig. 1D). Moreover, the expression level of CNPase, an alternative OL lineage marker, was similar between genotypes as determined by western blot (Fig. 1E and F). Thus, reduced Smn levels do not appear to affect the typical OL molecular maturation progression in vitro.

Figure 1

Oligodendrocyte differentiation and marker expression is unaltered upon Smn depletion. (A) Immunofluorescence micrographs of OL primary cultures from neonatal Smn^-/-;SMN2 or WT control mice after DIV3 and DIV6 of differentiation demonstrating the ability of both genotypes to progress through the molecular differentiation program. (B,C) Quantification of the percent of OLs expressing NG2, MAG and MBP at DIV3 and DIV6 for mutant and WT cultures. (D) Quantification of the intensity of MBP expression from DIV3 immunofluorescence images. (E) Fluorescence western blot analysis of CNPase and α-tubulin from primary OLs at DIV5. (F) Quantification of total CNPase normalized to α-tubulin (from blot pictured in E). Data represent the mean +/- standard error of the mean (SEM). Unpaired two-tailed Student’s t-test; n.s. indicates no significant difference; N = 3 (except for E and F where N = 4). Scale bar, 50 μm; applicable to all images.

To further interrogate the Smn^-/-;SMN2 OPC cultures, cell death and proliferation assays were conducted. The percentage of OPCs expressing Ki-67 was analysed as a marker of proliferation. We did not observe any significant difference in the expression of Ki-67 in differentiating OPCs after 4 h (Fig. 2A and B), nor at DIV3 (data not shown). Furthermore, we assessed the incidence of apoptotic cell death in these cultures using cleaved caspase 3 as a marker. There was no significant difference in the percentage of cleaved caspase 3-positive OLs between the genotypes at DIV3, suggesting that Smn reduction does not influence survival of OL-lineage cells in vitro (Fig. 2C and D). This correlates with the data above suggesting that Smn depletion does not impact OPC development.

Figure 2

The proportion of OPCs undergoing proliferation or of OLs undergoing cleaved caspase 3-mediated cell death is similar in wild type and Smn^-/-;SMN2 cultures. (A) Immunofluorescence micrographs of Smn^-/-;SMN2 or WT after 4 h of differentiation demonstrating similar expression of Ki67 in NG2-positive OPCs (arrowheads). (B) Quantification of the percent of OPCs that are double-positive for NG2 and Ki67, indicating proliferative precursor cells at 4 h of differentiation. (C) Immunofluorescence micrographs of Smn^-/-;SMN2 or WT OL cultures at DIV3 highlighting similar incidence of cleaved caspase 3 in MAG-positive OLs. (D) Quantification of the percent of OLs that are cleaved caspase 3-positive (arrowheads) at DIV3 in each genotype. Data represent the mean +/- SEM. Unpaired two-tailed Student’s t-test; n.s. indicates no significant difference; N = 3.

Smn depletion does not impair the migratory capacity of primary OPCs

Successful myelination by the mature OL is highly dependent on the migration of OPCs to the unmyelinated axon. As such, we sought to determine the impact of Smn depletion on the migration of OPCs. We took advantage of a previously described in vitro OPC aggregate (OPCA) migration assay developed for the primary murine system (17). Following seeding of individual OPCAs, OPCs were allowed to undergo radial migration for 24 h, at which point they were fixed for quantification. Our analysis revealed that the number of Smn^-/-;SMN2 OPCs capable of reaching each concentric ring, overlaid 100 μm apart, was comparable to wild type (Fig. 3A and B). The total number of OPCs migrated, irrespective of distance, can be used as a measure of the ability to initiate migration. There was no significant difference in the total number of OPCs migrated from the seeded aggregate between genotypes (Fig. 3C). Of note, the diameter of the initial OPCA at seeding time was also consistent (Fig. 3D). Taken together, these results indicate that reduced levels of Smn do not contribute to a migratory phenotype in OPCs, including initiation and extent of migration.

Figure 3

OPC migration in vitro is normal in Smn^-/-;SMN2-derived cultures. (A) Immunofluorescence micrographs of OPCAs from Smn^-/-;SMN2 or WT mice following 10 h of migration. (B) Quantification of the percentage of OPCs within rings placed 100, 200, 300 or 400 μm away from the original OPCA. There was no difference in the distances migrated by Smn^-/-;SMN2 OPCs at 10 h compared to WT. (C) Quantification of the total number of OPCs migrated from the OPCA, indicating migration initiation. Smn^-/-;SMN2 OPCs are equally capable of initiating migration as WT. (D) Sphere diameter of the initial OPCA at seeding time is similar in both genotypes. Data represent the mean +/- SEM. Repeated measures two-way ANOVA with Bonferroni post-tests (B) or unpaired two-tailed Student’s t-test (C); n.s. indicates no significant difference; N = 3. Scale bar, 50 μm; applicable to all images.

Reduction in Smn level does not inhibit morphological maturation of OLs in primary culture

A crucial component of OL development is the extension of a complex network of processes and the subsequent formation of membranous sheets. This transformation is collectively referred to as morphological maturation, and is required for efficient myelination. To quantify morphological maturation, primary OLs were binned into one of four stages as has been previously reported, based on their increasing complexity in vitro (Fig. 4) (18). At DIV3, the culture is expected to be predominantly stage 3 OLs, which have formed a highly complex network of processes and represent an intermediate OL. As differentiation continues to DIV6, OLs are predominantly stage 4, which are those producing a membranous sheet reminiscent of myelin. Stage 4 cells are considered to be morphologically mature OLs. Quantification of primary OLs using this staging system revealed no significant difference in morphological differentiation between Smn^-/-;SMN2 and wild type at DIV3 and DIV6. Importantly, the OLs derived from control animals exhibited the expected morphological complexities at each time point, such that the majority of OLs at DIV3 were highly branched (stage 3) with progression to a predominantly membranous culture (stage 4) at DIV6. Additionally, Sholl analysis was applied to OLs at DIV3. This measure quantifies the number of times a process of a given cell intersects with concentric rings digitally overlaid 10 μm apart. There was no significant difference between genotypes in the number of intersections across each of the ten rings, nor was a difference detected in the total number of process intersections (Fig. 4D).

Figure 4

Loss of Smn does not affect morphological differentiation of OLs. (A) Immunofluorescence micrographs depicting representative morphologies of OLs at DIV3 and DIV6 derived from Smn^-/-;SMN2 or WT mice. (B, C) A four-tiered staging scheme was employed to quantify the morphological differentiation of OLs in vitro. In this method, OLs with predominantly primary, secondary or tertiary branching were binned as stage 1, 2 or 3, respectively. OLs producing any amount of myelin membrane were binned as stage 4. A typical stage 3 OL is depicted in (A) at DIV3, while the DIV6 OLs shown in (A) are considered stage 4. Quantification revealed no difference in the morphology of Smn^-/-;SMN2 OLs relative to WT at either DIV3 (B) or DIV6 (C). Quantification of Sholl analysis completed on DIV3 OLs depicting the number of cell process intersections with concentric rings overlaid 10 μm apart (D). Quantification revealed no difference in the size or complexity of the cells when represented either by the number of intersections across each ring (i) or the sum of all intersections (ii). Data represent the mean +/- SEM. Unpaired two-tailed Student’s t-test or repeated measures two-way ANOVA with Bonferroni post-tests; n.s. indicates no significant difference; N = 3. DIV3 scale bar, 20 μm; DIV6 scale bar, 50 μm; applicable to all images within the given time point.

Smn depletion does not impact OL myelinating capacity in vitro

Smn depletion appears to have no impact on OPC development or migration, however the effect of Smn depletion on functional OL myelination remained to be addressed. To this end, we used an OL/dorsal root ganglion neuron (DRGN) co-culture system (19). The extent of myelination can be quantified in this assay by the co-localization of the myelin marker MAG with neurofilament, a neuron-specific marker. At DIV6 of co-culture, Smn^-/-;SMN2 OLs were capable of contacting and subsequently enwrapping neighbouring neurites with MBP ^+ ve membrane to the same extent as OLs derived from control littermates (Fig. 5A and B). Importantly, the neurite density in these cultures was consistent between genotypes, eliminating this as a confounding variable (Fig. 5C). Cumulatively, these results suggest that Smn depletion in the OL does not affect in vitro myelination.

Figure 5

Smn depletion does not affect the ability of OLs to myelinate in an OL/DRGN co-culture system. (A) Immunofluorescence micrographs of primary OLs (MAG-positive, green) derived from Smn^-/-;SMN2 or WT in culture with dorsal root ganglion neurons (DRGNs; neurofilament-positive, red) at DIV6. (B) Quantification of the ability of OLs to co-localize with DRGNs, used as a measure of the in vitro ability to ensheath axons with myelin membrane. Represented as an arbitrary measure of MAG and neurofilament co-localization, as determined using ImageJ software. (C) Neurite density does not differ between Smn^-/-;SMN2 and WT co-cultures. This indicates that the data was not confounded by differences in assay establishment. Data represent the mean +/- SEM. Unpaired two-tailed Student’s t-test; n.s. indicates no significant difference; N = 3. Scale bar, 50 μm; applicable to all images.

Smn depletion does not influence OL maturation in vivo

Analysis of OLs in vitro is essential to discern possible intrinsic developmental defects in an Smn-deficient context. However, the relevance of the OL to the overall SMA phenotype requires in vivo analysis of OL development and function. We assessed cortical white matter, specifically the corpus callosum, for the presence of mature (MBP-expressing) OLs. Quantification of the number of MBP-positive cells relative to the area of the corpus callosum indicated that the number of mature OLs in the white matter of Smn^-/-;SMN2 mice did not differ from wild type (Fig. 6). Therefore, we concluded that the ability to progress through the typical OL maturation gamut in vivo was not compromised by Smn depletion.

Figure 6

Smn^-/-;SMN2 mice have equivalent numbers of mature OLs in the corpus callosum as WT littermates. (A) Immunofluorescence micrographs depicting similar incidence of MBP-positive OLs in the corpus callosum of WT and Smn^-/-;SMN2 mice at P5. Boxed regions have been magnified. (B) Quantification of the number of MBP-positive OLs within the corpus callosum, and normalized to the area of the quantified region. Data represent the mean +/- SEM. Unpaired two-tailed Student’s t-test; n.s. indicates no significant difference; N = 3.

CNS myelination is unaffected in Smn^-/-;SMN2 mice

In addition to confirming the ability of OLs to mature and express the appropriate lineage markers in vivo, it is imperative to investigate the effects, if any, of Smn depletion on overall myelination in the CNS. We, therefore, assessed myelin ultrastructure in the spinal cords of postnatal day 5 (P5) mice. Qualitatively, white matter ultrastructure, as seen through electron microscopy, appeared similar between the Smn^-/-;SMN2 mice and wild type controls (Fig. 7A). G-ratios were calculated for each genotype as a quantitative measure of in vivo myelination. This measures the ratio between the inner and outer diameters of the myelin sheath, such that a g-ratio of 1 would represent a completely amyelinated axon and smaller values represent thicker myelin. The average g-ratios for spinal cord are remarkably similar between the two genotypes; Smn^-/-;SMN2 g-ratios averaged 0.787, while g-ratios for wild type controls averaged 0.786 (Fig. 7C). It is important to note that typical g-ratio values for murine spinal cord are reportedly 0.79, which is consistent with our findings (20). Finally, g-ratios plotted by the axon calibre from spinal cord tissue of P5 wild type and Smn^-/-;SMN2 mice also showed no difference upon linear regression analysis (Fig. 7D).

Figure 7

Normal myelination in the spinal cord and optic nerve of symptomatic Smn-deficient mice. Electron micrographs showing glial cells and many myelinated axons of different calibers in transverse sections of the white matter from the L1 spinal cords of wild type and severe Smn^-/-;SMN2 mice at P5 (A) as well as optic nerve from wild type and Smn^2B/- mice at P21 (B). Magnification 4,000x for all images and scale bar is 2 μm for all panels. (C) Quantification of the average g-ratio in spinal cord of P5 wild type and severe Smn^-/-;SMN2 mice. Data represent the mean +/- SEM. Unpaired two-tailed Student’s t-test; n.s. indicates no significant difference; N = 3. (D) G-ratios plotted by axon caliber from spinal cord tissue of P5 wild type and Smn^-/-;SMN2 mice. Data represent individual measurements, N = 153 and 137 for wild type and Smn^-/-;SMN2, respectively. Comparison non-significant by linear regression analysis.

To confirm that myelination in the spinal cord is representative of central myelination as a whole, we performed similar studies in the optic nerve. The optic nerve is myelinated much later in development than the spinal cord. The animals in the severe model of SMA do not survive past P6, at which point the optic nerve remains almost completely amyelinated. As a result, we investigated the ultrastructure of the optic nerve at P21 in a less severe model of SMA (Smn^2B/-) (21). Again, qualitative analysis revealed little difference between the affected animals and their wild type counterparts (Fig. 7B). Collectively, these results indicate that depletion of Smn does not have any impact on the developmental myelination program in the CNS, as represented by spinal cord and optic nerve in two different models of SMA.

Discussion

It is well established that the Smn protein is ubiquitously expressed with vital housekeeping roles in many cell types. In the CNS, Smn function extends beyond splicing and snRNP assembly to include β-actin mRNA transport and maintenance of neuronal growth cones (9,22,23). The necessity of Smn function to development is apparent as complete Smn knockout models exhibit pre-implantation defects and early embryonic lethality (7). It is also clear that reductions in Smn levels can lead to dramatic effects on life expectancy and motor function integrity, enforcing the fundamental role of this protein (3).

Although SMA is largely considered a disease of the motor neuron, the deleterious effect of Smn-depletion on other cell types is evident and gaining acceptance in the field. Defects in the development of muscle, a crucial component of the motor unit, have been reported in a variety of SMA models. Importantly, these defects are not simply a consequence of motor neuron atrophy, but are also due to intrinsic changes following Smn depletion (24,25). Moreover, in the peripheral nervous system of mouse models of SMA, myelination defects have been detected and can be attributed to defects in the developing Schwann cell (15,16). Exploration of the CNS in models of SMA has uncovered inherent abnormalities in cortical neuron development, as well as astrocyte reactivity and signalling (8,12,13,26). Furthermore, the characterization of the impact of Smn depletion has expanded to include non-neuronal systems. Notably, intrinsic pancreatic defects have been identified in the context of SMA, which result in impaired glucose metabolism not only in models of SMA, but also in Type I SMA patients (27,28). This clear trend towards the discovery of unique defects in multiple cell types contributing to SMA pathogenesis led to the hypothesis that Smn-depletion would affect OL development and subsequent myelination processes.

Through extensive investigation of OLs in a severe model of SMA, it was surprising to see that Smn reduction did not affect OL growth, maturation or function. This is contrary to our expectations and, to our knowledge, this apparent lack of a phenotype has not been seen in any other cell type studied to date. Previous studies have particularly underscored the importance of Smn function in developmental processes, which correlates well with high Smn expression in pre- and neonatal stages (29,30). Importantly, this is a crucial time for OL development and myelination (14). Thus, it was expected that Smn depletion would at least delay OL maturation, if not completely inhibit this process.

It is possible that the reduction of Smn in OLs has impacted an aspect of cell growth, maturation or function, but cannot be detected through the sensitivity of the selected assays. All assays were completed on primary OLs during differentiation. This holds the limitation that committed OLs are not highly proliferative cells (31). Therefore, detecting impacts on proliferation may be more difficult than in progenitor cells. Also, only cleaved caspase 3-mediated apoptosis was explored as a means of cell death and therefore it remains a possibility that reduced levels of Smn in the OL would increase the incidence of death occurring by an alternative mechanism. However, similar measures have been successfully employed to parse out defects in OL development following other genetic and signalling manipulations, such as deletion of integrin-linked kinase and inhibition of N-WASP and Fyn kinase (17,18,32–34). Therefore, it is likely that the chosen assessments and time points would be sufficient to detect a biologically relevant effect of Smn depletion on OLs.

An alternative explanation for the lack of change in OLs following Smn loss is the concept of Smn threshold (21). It is not only established that slight alterations to Smn levels can dramatically improve or worsen phenotype, but also that cell types are differentially susceptible to Smn depletion. It is widely accepted that motor neurons are highly sensitive to Smn depletion and exhibit phenotypic abnormalities at much higher levels of Smn than other tissue types. Moreover, differential susceptibility within the neuromuscular pathology has been well characterized. For example, it is clear that the motor units of abdominal muscles are more vulnerable to deficient Smn levels than those associated with cranial muscles (35). On the notion of differential susceptibility, this study does not have the capacity to discern subtle defects present in a subset of OLs, such as those responsible for myelinating lower motor neurons. Due to the limitations of the primary cell culture technique employed, conclusions regarding the impact of Smn depletion could only be made in OLs derived from the cortex. However, the assessment of myelin ultrastructure with electron microscopy was completed on sections from the lumbar region of the spinal cord. The lack of overt phenotype in this measure does not support the potential for vulnerable subpopulations of OLs in the lower regions of the spinal cord.

The seemingly insignificant impact of Smn depletion on OLs may also be explained by the induction of an alternative compensatory pathway, mitigating any developmental delays or defects. While this is improbable, as compensatory pathway activation has not been detected in similar essential cell types upon loss of Smn, it remains a possibility that should be further investigated. The discovery of a compensatory mechanism in OLs would be highly beneficial to the field, as manipulation of these pathways in affected cell types could provide a novel therapeutic avenue in SMA. Conversely, the molecular mechanisms mediating the effects of Smn depletion in other cell types may not be vital to OL development, thus explaining the seemingly normal growth of this cell type. However, this is unlikely as Smn reduction in other cell types impacts crucial pathways for OL growth, including the Rho family GTPases (36,37). Smn depletion correlates with increased RhoA activation in skeletal muscle, and inhibition of this activity has been sufficient to restore the expression of lineage markers and reverse growth defects (37).

In summary, we have taken advantage of a severe mouse model of SMA to perform a comprehensive characterization of OL functionality in vitro and in vivo upon Smn depletion. Our results indicate that unlike many neural cell types, Smn depletion does not impair proper development of OLs. Upon reduction of Smn, OLs remain capable of executing the complete differentiation program, of retaining migrating ability, and maintaining both in vitro and in vivo myelination capacity. These results have unveiled a novel system to exploit in the discovery of disease modifiers or compensatory mechanisms to further our understanding of differential susceptibility in SMA.

Materials and Methods

Mice

The mice used in this work were cared for according to the Canadian Council on Animal Care guidelines. Ethical approval for experiments conducted was obtained from the University of Ottawa Animal Care Committee under protocol number OHRI-1927. FVB Smn^+/-;SMN2 mice were obtained from Jackson Labs and were bred to generate the severe model of SMA (Smn^-/-;SMN2). For in vivo optic nerve analysis, a less severe model of SMA (Smn^2B/-) was used (21). Mice of either sex were used for all experiments.

Primary OL cell culture

OL cultures were established as described previously (19). Briefly, a mixed glial culture (MGC) was established from cortices removed from neonatal mice. The MGC was grown for 9–11 days, at which point purified OPCs were shaken off and seeded onto coverslips coated with human placental merosin (laminin-2 (Ln-2); Millipore) in either differentiation or migration media. OPCs were allowed to migrate for 10 h, or left to differentiate for three or six days (DIV3 and DIV6, respectively).

Migration assay

OL migration was assessed in vitro using the OPC aggregate migration protocol described previously (17). MGCs were established as above. Following the high speed shake purification, OLs form aggregates (OPCAs). These clusters of OLs were collected, seeded on Ln-2 coated coverslips and allowed to migrate for 10 h. Phase contrast images of OPCAs were taken at the time of seeding to record diameter. To quantify migration, exclusion zones 1.5x the seeding diameter of the OPCA were digitally overlaid using Photoshop. Only OPCs outside of this exclusion zone were counted. Concentric rings were then digitally overlaid with diameters increasing by increments of 100 μm (100 μm, 200 μm, 300 μm and 400 μm). Migration distance was quantified as the number of OPCs falling within each concentric ring and depicted as the percentage of OPCs in each ring. The total number of OPCs migrated was summated regardless of distance travelled and used as a measure of migration initiation.

In vitro myelination assay

OL/dorsal root ganglion neuron (DRGN) co-cultures were derived as described previously (19). Briefly, DRGs were dissected from the spinal cords of P5-P10 wild type mice. Isolated DRGNs were seeded on Ln-2 coated coverslips and cultured for nine days. Purified OPCs were then seeded on a dense neurite bed and allowed to differentiate for six days. Co-localized regions of MAG and neurofilament were measured using the integrated density plugin of ImageJ software. Neurite density was quantified as the mean gray value of the entire neurofilament signal in a given image.

Immunohistochemistry – cell culture

Coverslips were fixed with 100% methanol at 20 °C for 10 min or 3% paraformaldehyde (PFA) at room temperature for 15 min. Coverslips were then washed with PBS, permeabilized with 0.1% Triton X-100, blocked with 10% goat serum (GS), and incubated with primary antibodies against MAG (EMD Millipore), MBP (AbD Serotec), neurofilament-200 (NF-200; Sigma-Aldrich), cleaved caspase 3 (Asp175; Cell Signaling Technology), Ki67 (Leica Biosystems) and NG2 (EMD Millipore) in blocking solution at 4 °C overnight. OPCAs in the migration assay were additionally stained for F-actin, which included blocking for 30 min with 0.1% bovine serum albumin (BSA) and then incubation with rhodamine phalloidin (Life Technologies) in BSA for 45 min. Coverslips were then washed with PBS and incubated with Alexa Fluor-conjugated secondary antibodies (Alexa-488, Alexa-555, Alexa-647; Invitrogen) and/or rhodamine phalloidin (Invitrogen). Samples were counterstained with Hoechst nuclear stain and mounted in Dako mounting medium.

Immunofluorescent images were acquired using an Axio Imager M1 microscope with an AxioCamHR HRm Rev.2 camera and Axiovision 4.8.2 software. A Zeiss LSM 510 Meta DuoScan microscope (Zen 8.0 software) was employed for confocal microscopy. Phase contrast imaging was conducted using an Axiovert 200M microscope equipped with an AxioCamHR HRm Rev.2 camera and Axiovision 4.6 software.

Western blotting

At differentiation day 5 (DIV5), primary OL cultures were cooled on ice for 2 min and washed with ice-cold PBS. Cells were scraped off the dishes into 1x RIPA protein lysis buffer (Sigma) and centrifuged at high speed to remove insoluble material. Samples were separated by SDS-PAGE in a 15% gel. The membrane was incubated 1:1000 CNPase (Abcam) and 1:50000 alpha-tubulin (Cell Signaling) primary antibodies overnight at 4 °C in Odyssey blocking buffer (Li-Cor Biosciences) followed by a 1 hour room temperature incubation with secondary antibody (IRDye 680RD and 800CW; Li-Cor Biosciences) at 1:10000 in Odyssey blocking buffer. The blot was visualized and bands quantified using the Li-Cor Odyssey CLx Infared Imaging System.

Immunohistochemistry – murine tissue

Whole brain samples were dissected from P5 mice and fixed overnight in 4% PFA at 4 °C followed by overnight cryopreservation in 30% sucrose/PBS at 4 °C and then embedded in a 1:1 mixture of 30% sucrose/OCT (Sakura). Thirty μm coronal sections were cut and prepared for staining. Briefly, coronal sections were washed with PBS, permeabilized with 0.5% Triton-X-100 and blocked in 10% GS, 1% BSA and 0.2% Triton-X-100. Primary antibodies were diluted in blocking solution, and incubated overnight at 4 °C. Sections were then washed, incubated with Alexa Fluor conjugated secondary antibodies (Invitrogen), counterstained with Hoechst nuclear stain, and mounted in Dako mounting medium. Following image acquisition (5x magnification), the area of the corpus callosum was measured by manual tracing. MBP ^+ ve OLs were subsequently counted in each image, and normalized to the area of the corpus callosum. The number of MBP ^+ ve OLs represents an average from two sections per tissue sample.

Transmission electron microscopy

Vertebral columns from P5 wild type and Smn^-/-;SMN2 mice were dissected and fixed overnight at 4ºC in Karnovsky’s fixative (4% paraformaldehyde, 2% glutaraldehyde and 0.1 M sodium cacodylate in phosphate-buffered saline, pH 7.4). After fixation, the lateral funiculus of the white matter from each L1 spinal cord was cut under a stereomicroscope into a segment of less than 1 mm of length with a surgical blade. All segments were subsequently washed twice in 0.1 M sodium cacodylate buffer for 1 h and once overnight at room temperature. Segments were post-fixed with 1% osmium tetroxide in 0.1 M sodium cacodylate buffer for 1 h and subsequently dehydrated in an ascending concentration of alcohol. Samples were infiltrated gradually in 100% Spurr resin. Finally, segments were embedded in fresh liquid Spurr resin and polymerized at 70ºC overnight. Ultrathin sections (80 nm) from the white matter of the L1 spinal cord were cut with an ultramicrotome and collected onto 200-mesh copper grids. These sections were stained with 2% aqueous uranyl acetate and with Reynold’s lead citrate, then observed under a transmission electron microscope (Hitachi 7100) at 4,000x, 10,000x, 40,000x and 100,000x magnifications. Optic nerve samples taken from P21 Smn^2B/- and wild type mice were processed in a similar manner for electron microscopy. Approximately 100 electron micrographs of the white matter were qualitatively examined per genotype.

G-ratio measurement

Axon and total fibre diameter from wild type and Smn^-/-;SMN2 white matter from L1 spinal cords were measured from electron micrographs (4,000x magnification). All diameter measurements were traced manually using ImageJ software. A total of 135–155 myelinated axons were measured from 12 electron micrographs by genotype. G-ratios were calculated as the ratio of axon diameter relative to the total fibre diameter, representing the inner and outer limits of the myelin sheath, respectively.

Statistical analysis

Statistical analyses were performed using Prism 5/6 GraphPad software. All experiments were conducted with a sample size of N = 3 per genotype. For most experiments, two-tailed Student’s t tests were used for statistical analyses, with significance set at p < 0.05. Migration distance and Sholl analysis were analysed by repeated measures two-way ANOVA with Bonferroni post-tests. Linear regression analysis was used to analyse g-ratios by axon calibre.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

We would like to thank all members of the Kothary lab for helpful discussion and specifically Dr. Justin Boyer for assistance in study initiation. R.W.M., J-P.M. and M-O.D are all recipients of the Frederick Banting and Charles Best Canadian Institutes of Health Research Doctoral Award. S.C. is a recipient of the University of Ottawa CNMD Scholarship in Translational Research Award.

Conflict of Interest statement. None declared.

Funding

This work was supported by Cure SMA/Families of SMA Canada; Muscular Dystrophy Association (USA) (grant number 294568); Canadian Institutes of Health Research (CIHR) (grant number MOP-130279); and the E-Rare-2 program from the CIHR (grant number ERL-138414).

Author notes

†

These authors contributed equally to this work.