The geometric and spatial constraints of the microenvironment induce oligodendrocyte differentiation

Temporal Regulation of Oligodendrocyte Differentiation.

OPCs appear in the rat spinal cord as early as embryonic day (E)15. These precursor cells can be seen to increase in number as they migrate throughout the spinal cord (Fig. 1A). OPCs continue to proliferate until shortly after birth, at which time the number of OPCs becomes relatively constant (Fig. 1 A and C). Beginning around postnatal day (P)8, we observe a gradual decline in the number of OPCs (Fig. 1 A and C). This decrease in OPC numbers is followed by the appearance of differentiated oligodendrocytes, which are undetectable in the rat spinal cord until approximately one week postnatal (Fig. 1 B and C). These in vivo observations suggest that developing OPCs make a temporally synchronized transition from proliferating progenitors to differentiated oligodendrocytes. By using an in vitro coculture system involving sensory dorsal root ganglion (DRG) neurons and purified OPCs (2), we can consistently replicate this temporally specific pattern of differentiation. In our system, we find that after seeding a standard density of OPCs (200,000), the precursor cells undergo a fixed period of proliferation before differentiation. Differentiation begins approximately 10 days after the OPCs are seeded onto the neurons (Fig. 1 D and E). This sequence of events closely approximates the pattern of OPC development that we observe in the rat spinal cord. What mechanisms coordinate the initiation of oligodendrocyte differentiation that we see both in our cocultures and in vivo? In other words, how is the timing of oligodendrocyte differentiation synchronized between OPCs?

Induction of Differentiation Requires a Critical Density of OPCs.

It is possible that the global synchronicity of this cell fate decision can be explained by previous studies that suggest that OPCs differentiate based on the presence of an intrinsic timer (3). Essentially, OPCs are programmed to undergo a set period of division and then differentiate. Related OPCs possess highly similar programs and are coordinated in the timing of their differentiation, even when they are relocated to separate environments. It is important to note that the intrinsic timer was identified in studies performed at clonal density in the absence of other cell types (4, 5). In contrast, an OPC in our system is subject to the potential external influences of axons and neighboring OPCs. Can an intrinsic program explain the timing of oligodendrocyte differentiation in the presence of these extrinsic influences? To test this possibility, we seeded neurons with three different densities of OPCs (Fig. 1 F–H). If the initiation of differentiation is regulated by an intrinsic timer, then OPCs in all three experimental conditions should differentiate simultaneously. Instead, we find that the time at which differentiation is initiated depends on the density at which OPCs are seeded. OPCs seeded at high density (2 million OPCs) start to differentiate after only five days in culture (Fig. 1H), whereas OPCs seeded at our standard density (200,000 OPCs) and low density (20,000 OPCs) must be cultured for approximately two (Fig. 1G) and three weeks (Fig. 1F), respectively, before differentiation begins. These results suggest that an intrinsic timer is not responsible for the initiation of differentiation in our coculture system. Instead, it appears that a critical density of OPCs must be reached in order for differentiation to begin. This conclusion is supported by Western blot analysis (Fig. 1 F–H), showing the expression profiles of the OPC marker, platelet-derived growth factor receptor α (PDGFRα), and the oligodendrocyte marker, myelin basic protein (MBP). Regardless of the initial density of OPCs seeded, the expression of PDGFRα must reach a threshold level before MBP expression is detected. In addition, these findings demonstrate that in our cocultures, the extent of oligodendrocyte differentiation is not proportional to the number of OPCs initially seeded. Together, these studies support the hypothesis that extrinsic factors must regulate the onset of oligodendrocyte differentiation.

Differentiation and Myelination in the Absence of Dynamic Axonal Signaling.

It is likely that neurons serve as extrinsic regulators of OPC development. What role does the axon have in the process of oligodendrocyte differentiation? Similar to previous studies (6), we find that oligodendrocyte differentiation is not density dependent in the absence of axons (data not shown). These results suggest that interactions between axons and OPCs may be responsible for the density-dependent coordination of oligodendrocyte differentiation. To test this possibility, we treated our cocultures with either conditioned medium from axons or with purified axonal membranes (data not shown). Because these classic experiments failed to induce oligodendrocyte differentiation, we took an alternative approach to clarify the role of the axon. To eliminate dynamic changes in axonal signaling, we fixed axons with paraformaldehyde before seeding OPCs. We find that high-density OPCs seeded onto fixed axons differentiate with the same timing and robustness as OPCs on live axons (Fig. 2 A and B). To our surprise, the OPCs seeded onto fixed axons could also form compact myelin (Fig. 2 C–E). These results clearly indicate that dynamic interactions between oligodendrocytes and axons are not required for either differentiation or myelination.

Inhibition of Proliferation Is Not Sufficient to Induce Differentiation.

Could differentiation result from the presence of a permissive environment rather than from induction by an instructive cue? Our findings suggest that the onset of differentiation is somehow inextricably tied to the process of OPC proliferation. Perhaps differentiation is merely the result of inhibited proliferation that occurs once a critical density of OPCs on axons is reached. After all, it is plausible that achieving a high density of OPCs could inhibit proliferation by depleting available mitogens or through contact-mediated interactions. If this hypothesis is true, then an experimental inhibition of proliferation should induce oligodendrocyte differentiation. To test this possibility, we seeded our standard density of OPCs onto fixed axons [supporting information (SI) Fig. S1]. It is known that axons secrete OPC mitogens such as PDGF (7). In our cocultures, axonal fixation eliminates the secretion of these mitogenic factors, and greatly reduces the extent of OPC proliferation (Fig. S1). These findings are in line with previous studies demonstrating that axonal fixation selectively halts proliferation, but does not affect the viability of glial cells (8). Despite the inhibition of proliferation in our fixed axon cocultures, the OPCs failed to differentiate in advance of OPCs on live axons. In fact, even after 20 days, very few differentiated oligodendrocytes are detectable in these fixed axon cocultures (Fig. S1). These results suggest that the absence of mitogenic factors prevents OPCs from reaching the critical density required for differentiation. Also, we tested whether it was possible to delay differentiation by increasing the amount of available mitogens. The addition of exogenous PDGF to live axon cultures does not inhibit OPC differentiation and may actually advance the differentiation process (Fig. S2). Taken together, these results demonstrate that the induction of differentiation at a critical density of OPCs is not the result of an inhibition of proliferation.

Spatial and Geometric Constraints Induce Differentiation.

Why do OPCs need to reach a critical density to differentiate? Quantification of the critical density suggests that differentiation is induced once OPCs reach a density of ≈500–600 PDGFRα⁺ cells per millimeter squared. This density does not depend on the number of OPCs initially seeded and is highly conserved between live and fixed axon cultures (Fig. 2 F–I). Also, quantification of MBP⁺ cells suggests that the number of differentiated oligodendrocytes is not proportional to the number of OPCs initially seeded. These results confirm that a significant amount of differentiation is induced only after a critical density of OPCs is reached. Is it possible that the induction of differentiation at a critical density requires density-dependent interactions between OPCs? To test this possibility, we attempted to induce differentiation in cocultures containing both OPCs and Schwann cells seeded onto live axons. When our standard density of OPCs is seeded onto neurons either alone (Fig. 1 E and G) or in the presence of a low density of Schwann cells (Fig. 3A), we fail to see any differentiated oligodendrocytes after five days in culture. However, when seeded with a high density of Schwann cells, we see a significant amount of differentiation after five days (Fig. 3B). These findings suggest that differentiation is not controlled by the expression of a signal exclusive to OPCs. Instead, the addition of a high density of Schwann cells is sufficient to mimic the critical density required for the initiation of differentiation.

Because Schwann cells and OPCs originate from distinct cell lineages and do not interact in vivo, it is not clear why Schwann cells would express a signal that induces OPCs to differentiate. Therefore, we propose an alternative explanation for the ability of Schwann cells to induce oligodendroycte differentiation. We hypothesize that an OPC differentiates not because of instructive intercellular signaling, but instead because of a restriction in axonal space proximal to the OPC. To test this hypothesis, Schwann cells were fixed and seeded onto live axons along with a standard density of OPCs (Fig. S3). When seeded at a high density, fixed Schwann cells were sufficient to induce oligodendrocyte differentiation after only five days. These results suggest that secreted factors do not control the ability of Schwann cells to induce differentiation, adding support to the hypothesis that differentiation is induced through a restriction of available axonal space. To rule out the possibility that a membrane-bound factor is required to induce oligodendrocyte differentiation, we conjugated axons with polystyrene beads measuring 20 μm in diameter, which is comparable in size to both OPCs and Schwann cells (Fig. 3C). Coating the beads with an antibody against an axonal protein allowed us to bind the beads at low and high density along the surface of live axons (Fig. 3C). When a standard density of OPCs was seeded onto neurons with a low density of beads, no oligodendrocytes were observed after 5 days in culture (Fig. 3D). However, we find that an increase in the density of beads along axons is sufficient to induce differentiation on either live (Fig. 3E) or fixed axons (Fig. S4). These findings indicate that the induction of differentiation at a critical density of OPCs depends on a restriction of available axonal space. However, in the absence of axons, the polystyrene beads do not influence differentiation (data not shown). Also, we find that the induction of differentiation depends not only on the density of beads along an axon, but also on their dimensions. Either an increase or decrease in the diameter of the beads eliminates the potential to induce differentiation (Fig. 4). These results suggest that the induction of oligodendrocyte differentiation depends on both the spatial and geometric parameters of the surrounding microenvironment.

Discussion

Our results suggest that physical characteristics of the microenvironment can influence cell fate decisions. We hypothesize that spatial and geometric constraints along an axon may induce differentiation through lateral compression of OPCs. How does a mechanical stimulus promote a change in cell fate? One possibility is that contractile forces can physically alter the size or shape of an OPC. Changes in cell shape have previously been shown to regulate processes such as cell survival (9) and cell fate decisions (10). It has been suggested that changes in cell shape may facilitate structural rearrangement within the cell (11). This intracellular reorganization could initiate interactions between upstream mediators and downstream effectors that are responsible for the induction of differentiation (12). Also, it has been demonstrated that changes in cell shape can be directly linked to structural changes in the nucleus (13). In our system, we observe that an increase in cell density correlates with a decrease in nuclear size, as shown by staining with DAPI. It is possible that changes to the size and structure of the nucleus induce transcriptional activity necessary for oligodendrocyte differentiation. Perhaps an increase in cell density affects transcription factors known to regulate differentiation, such as Nkx2.2, Sox10, Olig2, and the recently identified Yin Yang 1 (14–17). Understanding the relationship between intrinsic and extrinsic factors could help identify environmental conditions that will promote oligodendrocyte differentiation in the presence of chronic demyelination.

Treatment of demyelinating conditions may also require elucidation of the relationship between the various environmental stimuli that have been shown to regulate differentiation. In our studies we find that the presence of the axon is required for the induction of differentiation through geometric and spatial constraints. Our results are in line with previous studies, which suggest that the nature of the cellular substrate can influence the effect of a mechanical stimulus on cell fate decisions (11). For instance, the fate of mesenchymal stem cells can be modulated by the flexibility of the substrate on which the cells are cultured (18). It is possible that biophysical characteristics of the axon, such as its size, shape, and tensile strength could be instrumental in the density-dependent induction of differentiation. After all, the caliber of the axon has previously been shown to influence multiple aspects of the myelination process. Studies suggest that axon diameter can regulate whether axons are myelinated (19), the thickness of the myelin membrane, and the distribution of internodes along an axon (20). However, recent studies suggest that in the peripheral nervous system the regulation of myelin sheath thickness may be due solely to the fact that an increase in axon diameter correlates with an increase in the expression of the membrane-bound axonal factor neuregulin 1 (NRG1) type III (21, 22). As illustrated by these findings, it is important to determine whether it is biophysical or biochemical attributes (or a combination of both) that make the axon a necessary component of density-dependent differentiation.

It is possible that the induction of differentiation through environmental constraints depends not on the structural dimensions of the axon but, instead, on the expression of membrane-bound axonal signals. Perhaps these signals facilitate the ability of an OPC to transduce mechanical forces into the biochemical initiation of oligodendrocyte differentiation. This type of context-dependent regulation could represent an alternative explanation for the role of an axonal substrate in the density-dependent induction of oligodendrocyte differentiation. It is possible that the mechanical induction of differentiation depends on the formation of an adherens junction between the axon and the OPC. This theory is supported by studies that implicate ECM receptors and cell–cell adhesion molecules as key mediators of mechanotransduction (23). Interestingly, adhesion molecules have previously been shown to have a role in oligodendrocyte differentiation and myelination (24–28). ECM receptors such as integrins are also known to modulate the effects of extrinsic growth factors on various aspects of oligodendrocyte development (29, 30). Therefore, it is plausible that the neuronal-glial interaction facilitates the formation of an adhesion complex that is responsible for the integration of various chemical and mechanical factors regulating differentiation (31). This hypothesis could help explain how the axon can regulate differentiation through the expression of putative factors such as Jagged-1 (32, 33) and Lingo-1 (34, 35), as well as through its role in mechanotransduction. Because novel differentiation factors continue to be identified, it is essential to understand how a multitude of diverse extrinsic factors can be properly synthesized in the coordination of a single cell fate decision. Identifying the mechanisms responsible for the integration of extrinsic signals may be crucial for the establishment of an environment promoting remyelination.

Materials and Methods

Supplementary Material