ERK1/ERK2 MAPK signaling is required to increase myelin thickness independent of oligodendrocyte differentiation and initiation of myelination

Abstract

Wrapping of the myelin sheath around axons by oligodendrocytes is critical for the rapid conduction of electrical signals required for the normal functioning of the CNS. Myelination is a multistep process where oligodendrocytes progress through a well coordinated differentiation program regulated by multiple extracellular growth and differentiation signals. The intracellular transduction of the extracellular signals that regulate myelination is poorly understood. Here we demonstrate a critical role for two important signaling molecules, extracelluar signal-regulated protein kinases 1 and 2 (ERK1/ERK2), downstream mediators of mitogen-activated protein kinases, in the control of CNS myelin thickness. We generated and analyzed two lines of mice lacking both ERK1/ERK2 function specifically in oligodendrocyte-lineage cells. In the absence of ERK1/ERK2 signaling NG2⁺ oligodendrocyte progenitor cells proliferated and differentiated on schedule. Mutant oligodendrocytes also ensheathed axons normally and made a few wraps of compact myelin. However, the subsequent increase in myelination that correlated myelin thickness in proportion to the axon caliber failed to occur. Furthermore, although the numbers of differentiated oligodendrocytes in the adult mutants were unchanged, they showed an inability to upregulate the transcription of major myelin genes that normally occurs during active myelination. Similarly, in vitro ERK1/ERK2-deficient oligodendrocytes differentiated normally but failed to form typical myelin-like membrane sheets. None of these effects were observed in single ERK1 or ERK2 mutants. These studies suggest that the predominant role of ERK1/ERK2 signaling in vivo is in promoting rapid myelin growth to increase its thickness, subsequent to oligodendrocyte differentiation and the initiation of myelination.

Figure 1.

Myelin sheath thickness is reduced in Erk1/Erk2 dKO. A, Transverse sections of cervical spinal cord from Erk1/Erk2-CNP^Cre dKO and littermate control mice immunolabeled for MBP (green) and NF-M (red) at P14 and P30 show that the area of the ventrolateral white matter is significantly decreased in dKO compared with littermate control at both ages (P14: control = 0.52 ± 0.04 mm², dKO = 0.38 ± 0.02 mm², **p < 0.01, N = 3; P30: control = 0.91 ± 0.06 mm², dKO = 0.73 ± 0.06 mm², *p < 0.05, SEM, N = 3–4). High-magnification images (bottom) of P30 spinal cord show less MBP⁺ myelin around axons in the dKO. Scale bars: 200 and 10 μm. B, EM micrographs of ventral spinal cords at P16 and P30 show that dKO axons are wrapped by thinner layers of myelin, compared with littermate controls. High-magnification images show that the periodicity of myelin is normal in the dKOs. Scale bars: 2 and 0.1 μm. C, Quantification of g-ratios of individual fibers in relation to respective axon diameters (presented as scatter plots) at P16 and P28/P30 show decreased myelin thickness in Erk1/Erk2 dKO (pink circles) compared with littermate controls (blue circles). D, Quantification presented as myelin thickness (μm) in relation to respective axon diameters shows that in the controls myelin thickness increases in parallel to axon diameter but in the dKO axons of all diameters are wrapped by myelin of similar thickness, which is disproportionately thinner than normal. Approximately 100–400 axons each were measured from two controls and two dKOs at P16 and two controls and two dKOs pooled from P28 and P30.

Figure 2.

Proliferation and numbers of OPCs are not affected in Erk1/Erk2 dKO. A, Transverse sections of P0 cervical spinal cord from Erk1/Erk2-NG2^Cre or Erk1/Erk2-CNP^Cre and littermate control mice were immunolabeled for Olig2 (green) and Ki67 (red) to identify proliferating OPCs. Inset, The enlarged view of the boxed region shows double-labeled cells (yellow). B, The total numbers of Olig2⁺;Ki67⁺ double-labeled cells (expressed as a percentage of Olig2⁺ cells) and the density of Olig2 cells in the whole spinal cord sections of Erk1/Erk2-NG2^Cre or Erk1/Erk2-CNP^Cre mice did not differ from that of littermate controls. Scale bar, 100 μm. Three sections each from three mice of each genotype were analyzed. Error bars indicate SEM, N = 3.

Figure 3.

Differentiation of oligodendrocytes and the initiation of myelination are not affected in Erk1/Erk2 dKO. A, Transverse sections of P0 and P1 cervical spinal cord from Erk1/Erk2-NG2^Cre and control mice were analyzed by in situ hybridization for PLP mRNA to mark differentiated oligodendrocytes. Quantification of PLP⁺ oligodendrocytes in half-sections of P0 spinal cord shows no difference in their numbers. Three sections each from three mice of each genotype were analyzed. Error bars indicate SEM, N = 3. Representative images of PLP mRNA⁺ oligodendrocytes at P1 are shown. Scale bar, 100 μm. B, Explant cultures initiated from individual spinal cords of P2 Erk1/Erk2-NG2^Cre and control mice, double-immunolabeled for immature oligodendrocyte marker O1 (red) and total oligodendrocyte-lineage cell marker Olig2 (green) at 4 DIC, show that the percentage of O1⁺/Olig2⁺ oligodendrocytes in the mutants are not significantly different from the controls (control = 20.5 ± 1.9; dKO = 16.3 ± 0.8, SEM, p = 0.05, N = 3 animals per genotype). C, Dissociated cultures initiated from individual spinal cords of P2 Erk1/Erk2-CNP^Cre mutant and littermate controls were double-immunolabeled at 5 DIC with O4, marker of immature oligodendrocytes (HPC7 antibody which labels GalC⁺ or O1⁺ stage), and MBP, marker of mature oligodendrocytes. The percentage of total oligodendrocyte-lineage cells (O4⁺) that differentiated into oligodendrocytes (at both stages of maturation), are comparable in mutant and control cultures. Error bars indicate SEM (p = 0.16; N = 3 animals per genotype). D, Transverse sections of spinal cords from P0 Erk1/Erk2-CNP^Cre, P1 Erk1/Erk2-NG2^Cre dKOs, and littermate controls double-immunolabeled for MBP (green) and NF-M (red) show similar patterns of MBP-immunolabeled myelinated fibers, indicating similar onset of MBP expression and myelin biogenesis in the dKOs. Representative images out of 3–8 controls and 3–5 dKOs are shown. E, Representative EM images of ventral spinal cords out of three Erk1/Erk2-CNP^Cre dKO and three control mice at P4 show a similar pattern of axonal ensheathment by thin myelin sheaths. Quantification from ∼1000 axons from four 3000× EM images of each genotype show similar percentages of myelinated or unmyelinated axons in the Erk1/Erk2-CNP^Cre dKOs and littermate controls, together indicating normal onset of axonal ensheathment and initiation of myelin wrapping in dKOs.

Figure 4.

The levels of PLP mRNA and MBP mRNA, but not the numbers of oligodendrocytes, are reduced in Erk1/Erk2 dKO. Transverse sections of P14 and P30 cervical spinal cords from control and Erk1/Erk2-CNP^Cre dKO mice were analyzed by in situ hybridization for PLP (A) or MBP (C) mRNA expression. A, Quantification of total numbers of PLP mRNA⁺ oligodendrocytes in the lateral ventral white matter (WM) from half-sections of the spinal cords shows no difference in the numbers of oligodendrocytes between control and dKO at either of the ages. The PLP mRNA signal intensity per oligodendrocyte is, however, reduced in dKO compared with controls (boxed areas of P30 are shown at higher magnifications in the bottom). Three sections each from three mice of each genotype at each age were analyzed. Error bars indicate SEM, N = 3. Scale bars: 100 and 50 μm. B, Immunoblot analysis of P20 cervical spinal cord homogenates (0.5 μg total protein per lane) from three mice of each genotype shows a decrease in the levels of PLP protein in the Erk1/Erk2-CNP^Cre dKO compared with controls. β-Actin is shown as a loading control. C, The expression of MBP mRNA is reduced in Erk1/Erk2-CNP^Cre dKOs compared with controls at P14 and is almost completely abolished from the fibers by P30 but remains in oligodendrocyte cell bodies in dKO. Representative images are shown from three Erk1/Erk2-CNP^Cre dKOs and 4–5 controls for each group. Expression of MBP mRNA is also reduced in Erk1/Erk2-NG2^Cre dKO compared with littermate controls. D, Quantification of mRNA levels by qRT-PCR shows a significant reduction in the expression of PLP and MBP transcripts in the P20 mutant spinal cords compared with controls. Error bars indicate SEM. **p < 0.01, N = 3–6.

Figure 5.

Reduction in myelin gene expression and myelin sheath thickness occurs in other regions of the CNS of Erk1/Erk2 dKOs. A, B, Sagittal sections of P30 hindbrain from Erk1/Erk2-CNP^Cre dKO and littermate control mice, analyzed by in situ hybridization for the expression of MBP mRNA (A) and PLP mRNA (B), show that MBP mRNA signal intensity is reduced in the cerebellum and brainstem of the mutants compared with controls (A). Similarly, the level of PLP mRNA expression by oligodendrocytes is reduced in the cerebellum of the mutants compared with controls (B). Bottom (B) shows higher magnification images of the boxed area. Scale bars: 200 μm. Multiple sections from two mice of each genotype were analyzed and representative images from matched sections are shown. C, EM micrographs of cerebellar peduncles at P30 show that dKO axons are wrapped by thinner myelin sheaths, compared with littermate controls. Scale bar, 1 μm. D, Scatter plots of g-ratios and myelin thickness (μm) measurements of 200–400 axons from two mice of each genotype at P30 show statistically significant decrease in myelin thickness in the cerebellum of Erk1/Erk2-CNP^Cre dKO (pink circles) compared with littermate controls (blue circles) (average g-ratio, p = 0.5 × 10⁻¹³).

Figure 6.

Membrane formation by oligodendrocytes in vitro is inhibited in Erk1/Erk2 dKO. Dissociated primary culture from individual spinal cords of P2 controls (n = 6) and Erk1/Erk2-CNP^Cre (n = 7) pups were grown in duplicate in defined media and analyzed at 5, 8, and 13 DIC by immunolabeling for MBP (red) and for the nuclear stain Hoechst (green) to examine the morphology of differentiated oligodendrocytes. Initially, at 5 DIC control (a,c) and dKO oligodendrocytes (b,d) form similar process networks. At 13 DIC, while the control oligodendrocytes have extended large membrane sheaths (e,g), the majority of dKO oligodendrocytes are unable to make membranes but remain viable as evident by the noncondensed nucleus of these cells (f,h, arrowheads). For quantification of the size of the membranes, a total of 100–150 oligodendrocytes from three independent representative cultures of each genotype were measured at each time point (8 and 13 DIC) (i). To count the numbers of oligodendrocytes that made membrane sheaths, 200–300 MBP⁺ oligodendrocytes were examined from three independent control and three dKO cultures each (j). Note that membranes made by mutant oligodendrocytes are significantly smaller than controls and fewer mutant MBP⁺ oligodendrocytes make membrane sheaths compared with control oligodendrocytes. Arrowheads point to the cell bodies of the oligodendrocytes. Scale bar, 50 μm. Error bars indicate SEM. *p < 0.05, **p < 0.01, N = 3.