Rnd2 differentially regulates oligodendrocyte myelination at different developmental periods

Abstract

In the CNS, oligodendrocyte precursor cells differentiate into oligodendrocytes to wrap their plasma membranes around neuronal axons, generating mature neural networks with myelin sheaths according to spatial and temporal patterns. While myelination is known to be one of the most dynamic cell morphological changes, the overall intrinsic and extrinsic molecular cues controlling myelination remain to be fully clarified. Here, we describe the biphasic roles of Rnd2, an atypical branch of the Rho family GTPase, in oligodendrocyte myelination during development and after maturation in mice. Compared with littermate controls, oligodendrocyte-specific Rnd2 knockout mice exhibit decreased myelin thickness at the onset of myelination but increased myelin thickness in the later period. Larger proportions of Rho kinase and its substrate Mbs, the signaling unit that negatively regulates oligodendrocyte myelination, are phosphorylated at the onset of myelination, while their smaller proportions are phosphorylated in the later period. In addition, we confirm the biphasic role of Rnd2 through experiments with oligodendrocyte-specific Rnd2 transgenic mice. We conclude that Rnd2 positively regulates myelination in the early myelinating period and negatively regulates myelination in the later period. This unique modulator thus plays different roles depending on the myelination period.

FIGURE 1:

Changes in expression levels of Rnd2 and myelin marker proteins. (A) Whole brain tissues from postnatal days 1 to 21 (P1 to P21) and adult stages were isolated, lysed, and used for immunoblotting with the respective antibodies for Rnd2, CNPase, MBP, and control actin. (B) Rat oligodendrocytes and dorsal root ganglion (DRG) neurons were cocultured for 1–21 d in vitro (DIV), lysed, and used for immunoblotting with the respective antibodies for Rnd2, CNPase, MBP, and actin. (C) Oligodendrocytes were lysed and used for immunoblotting with the respective antibodies for Rnd2, CNPase, MBP, and actin. (D) Cross-sections of 7-, 14-, or 21-d-old mouse corpus callosum were costained with antibodies against Rnd2 (green) and PDGFRα, O4, or CC1 (red), respectively. Data are representative of multiple iterations. The scale bars indicate 50 μm. (E) The percentage of PDGFRα⁺, O4⁺, or CC1⁺ cells among the Rnd2⁺ cells is shown (n = 8–20 of total sections stained and counted from three sets of independent littermate mice). (F) Longitudinal sections of 14-d-old mouse corpus callosum were costained with antibodies against Rnd2 (green) and NF (red). Data are representative of multiple iterations. The scale bars indicate 50 μm.

FIGURE 2:

Mbp-Cre–mediated Rnd2 knockout mice exhibit decreased myelin marker protein expression at postnatal day 14. (A) Oligodendrocyte-specific MBP promoter–driven Cre recombinase transgenic mice were used for deletion of the loxP-flanked exons of the rnd2 gene. The knockout cassette as well as the primer and probe positions are shown. (B) Antibodies against MBP (red) and Olig2 (green) were used for costaining in 14-d-old Mbp-Cre–driven Rnd2 conditional knockout (Rnd2^fl/fl; Mbp-Cre) or control mouse brain cross-sections. The scale bars indicate 200 μm. (C) The intensity of MBP staining per square millimeter was relatively semiquantified (**, p < 0.01 in Student’s t test; n = 6–8 of total sections stained and counted from three sets of independent littermate mice). (D) The Olig2⁺ cells per square millimeter were counted (n = 9–12 of total sections stained and counted from three sets of independent littermate mice). (E) Brain cross-sections of conditional knockout (Rnd2^fl/fl; Mbp-Cre) and control mice at postnatal day 14 were costained with an anti-CNPase antibody (red) and Olig2 (green). The scale bars indicate 200 μm. (F) The intensity of CNPase staining per square millimeter was relatively semiquantified (**, p < 0.01 in Student’s t test; n = 6 of total sections stained and counted from three sets of independent littermate mice). (G) Tissue lysates from 14-d-old conditional knockout or control (Ctrl) mouse whole brains were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (H) CNPase expression levels are shown statistically (**, p < 0.01 in Student’s t test; n = 4 blots from four sets of independent littermate mice).

FIGURE 3:

Mbp-Cre–mediated Rnd2 knockout mice exhibit decreased myelin thickness at postnatal day 15. (A) Electron microscopic images of the corpus callosum of conditional knockout (Rnd2^fl/fl; Mbp-Cre) and control mice at postnatal day 15 are shown. The scale bars indicate 2 μm. (B) Graph of g-ratios of myelinated axons for axon diameters is shown (n = 343 axons for three different knockout mice and n = 297 axons for their littermate controls). (C) Average g-ratios are shown (**, p < 0.01 in Student’s t test). (D) Graph of the distributions of g-ratios is shown. (**, p < 0.01, *, p < 0.05 in one-way ANOVA).

FIGURE 4:

Knockdown of Rnd2 decreases myelin formation in cocultures at 21 d. (A) Knockdown efficiencies of two nonoverlapping shRNAs (Rnd2#1 and Rnd2#2) were confirmed by immunoblotting. (B) Oligodendrocytes that had been transfected with Rnd2 shRNAs or control cells were cocultured with neurons for 21 d and then costained with antibodies against MBP (red), NF (green), and DAPI (blue) in order to detect myelin segments in regions of interest. The scale bars indicate 200 μm. Magnified views of the myelinated oligodendrocytes in the dotted white squares are also shown. The scale bars indicate 50 μm. (C) Oligodendrocytes that had been transfected with Rnd2 shRNAs or control cells were cocultured with neurons for 21 d and then costained with antibodies against MBP (red) and Olig2 (green). The scale bars indicate 200 μm. Magnified views of the myelinated oligodendrocytes in the dotted white squares are also shown. The scale bars indicate 50 μm. (D) The percentage of MBP⁺ cells among the Olig2⁺ cells per field (**, p < 0.01 in one-way ANOVA; n = 8–12 fields of three independent experiments). (E) The myelin segments per field were counted (**, p < 0.01 in one-way ANOVA; n = 10–11 fields of three independent experiments). (F) The total sheath length of myelinated oligodendrocytes per field was measured (**, p < 0.01 in one-way ANOVA; n = 15 fields of three independent experiments). (G) The number of Olig2⁺ cells per field is shown (n = 8–12 fields of three independent experiments). (H) The lysates were immunoblotted with an antibody against MBP, CNPase, or actin. (I) CNPase expression levels are shown statistically (**, p < 0.01 in Student’s t test; n = 4 blots of four independent experiments).

FIGURE 5:

Mbp-Cre–mediated Rnd2 knockout mice exhibit increased myelin marker protein expression and myelin thickness at postnatal day 28. (A) Brain coronal sections of conditional knockout (Rnd2^fl/fl; Mbp-Cre) and control mice at postnatal day 28 were costained with an anti-MBP antibody (red) and Olig2 (green). The scale bars indicate 200 μm. (B) The intensity of MBP staining per square millimeter was relatively semiquantified (**, p < 0.01 in Student’s t test; n = 10 of total sections stained and counted from three sets of independent littermate mice). (C) The Olig2⁺ cells per square millimeter were counted (n = 6–10 of total sections stained and counted from three sets of independent littermate mice). (D) Tissue lysates from 28-d-old brains were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (E) CNPase expression levels are shown statistically (*, p < 0.05 in Student’s t test; n = 4 blots from four sets of independent littermate mice). (F) Electron microscopic images of the corpus callosum of conditional knockout (Rnd2^fl/fl; Mbp-Cre) and control mice at postnatal day 28 are shown. The scale bars indicate 2 μm. (G) Graph of g-ratios of myelinated axons for axon diameters is shown (n = 514 axons for three different knockout mice and n = 504 axons for their littermate controls). (H) Average g-ratios are shown (**, p < 0.01 in Student’s t test). (I) Graph of the distributions of g-ratios is shown (**, p < 0.01, *, p < 0.05 in one-way ANOVA). (J) Electron microscopic images of the corpus callosum of conditional knockout and control mice at 2 mo are shown. The scale bars indicate 2 μm. (K) Graph of the g-ratios of myelinated axons for axon diameters is shown (n = 297 axons for three different knockout mice and n = 287 axons for their littermate controls). (L) Average g-ratios are shown (**, p < 0.01 in Student’s t test). (M) Graph of the distributions of g-ratios is shown (**, p < 0.01, *, p < 0.05 in one-way ANOVA).

FIGURE 6:

Knockdown of Rnd2 increases myelin formation in cocultures at 28 d. (A) Oligodendrocytes that had been transfected with Rnd2 shRNAs or control cells were cocultured with neurons for 28 d and then costained with antibodies against MBP (red), NF (green), and DAPI (blue). The scale bars indicate 200 μm. Magnified views of the myelinated oligodendrocytes in the dotted white squares are also shown. The scale bars indicate 50 μm. (B) Oligodendrocytes that had been transfected with Rnd2 shRNAs or control cells were cocultured with neurons for 28 d and then costained with antibodies against MBP (red) and Olig2 (green). The scale bars indicate 200 μm. Magnified views of the myelinated oligodendrocytes in the dotted white squares are also shown. The scale bars indicate 50 μm. (C) The percentage of MBP⁺ cells among the Olig2⁺ cells per field (**, p < 0.01, *, p < 0.05 in one-way ANOVA; n = 6–7 fields of three independent experiments). (D) The myelin segments per field were counted (**, p < 0.01 in one-way ANOVA; n = 17–25 fields of three independent experiments). (E) The total sheath length of myelinated oligodendrocytes per field was measured (**, p < 0.01 in one-way ANOVA; n = 22 fields of three independent experiments). (F) The number of Olig2⁺ cells per field is shown (n = 6–9 fields of three independent experiments). (G) The lysates were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (H) CNPase expression levels are shown statistically (**, p < 0.01 in Student’s t test; n = 5 blots from four independent experiments).

FIGURE 7:

Rnd2 transgenic mice exhibit increased myelin marker protein expression and myelin thickness at postnatal day 14, whereas both are decreased in transgenic mice at postnatal day 28. (A) Tissue lysates from 14-d-old Rnd2 transgenic (Rnd2^Tg) and control mouse brains were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (B) MBP expression levels are shown statistically (*, p < 0.05 in Student’s t test; n = 4 blots from four different sets of littermate mice) (C) CNPase expression levels are shown statistically (**, p < 0.01 in Student’s t test; n = 5 blots from five different sets of littermates). (D) Brain cross-sections of Rnd2 transgenic (Rnd2^Tg) and control mice at postnatal day 14 were costained with an anti-CNPase antibody (red) and Olig2 (green). The scale bars indicate 200 μm. (E) The intensity of CNPase staining per square millimeter was relatively semiquantified (*, p < 0.05 in Student’s t test; n = 5 sections from three different sets of littermates). (F) The Olig2⁺ cells per square millimeter were counted (n = 9–12 sections from three different sets of littermates). (G) Electron microscopic data of 14-d-old corpus callosum are shown. The scale bars indicate 2 μm. (H) Graph of g-ratios of myelinated axons for axon diameters is shown (n = 289 axons for three different transgenic mice and n = 273 axons for their littermate controls). (I) Average g-ratios are shown (**, p < 0.01 in Student’s t test). (J) Graph of the distributions of g-ratios is shown (**, p < 0.01, *, p < 0.05 in one-way ANOVA). (K) Tissue lysates from 28-d-old Rnd2 transgenic and control (Ctrl) mouse brains were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (L) MBP expression levels at 28 d are shown statistically (**, p < 0.01 in Student’s t test; n = 3 blots from three different sets of littermates). (M) CNPase expression levels at 28 d are shown statistically (**, p < 0.01 in Student’s t test; n = 3 blots from three different sets of littermates). (N) Average g-ratios (n = 309 axons for three different transgenic mice and n = 291 axons for their littermate controls) at 28 d are shown (**, p < 0.01 in Student’s t test). (O) Graph of the distributions of g-ratios is shown (**, p < 0.01, *, p < 0.05 in one-way ANOVA).

FIGURE 8:

Mbp-Cre–mediated Rnd2 knockout mice exhibit increased Rho kinase signaling at postnatal day 14 and decreased Rho kinase signaling at postnatal day 28. (A) Tissue lysates from 14-d-old conditional knockout (Rnd2^fl/fl; Mbp-Cre) and control mice were used for immunoblotting with the respective antibodies for phosphorylated Mbs (pMbs), Mbs, phosphorylated Rho kinase (pRho kinase), and Rho kinase are shown. (B, C) Statistical data of phosphorylated Mbs and phosphorylated Rho kinase (**, p < 0.01, *, p < 0.05 in Student’s t test; n = 4 blots from four different sets of littermates). (D) Tissue lysates from 28-d-old conditional knockout (Rnd2^fl/fl; Mbp-Cre) and control mice were used for immunoblotting with the respective antibodies for pMbs, Mbs, pRho kinase, and Rho kinase and are shown. (E, F) Statistical data of phosphorylated Mbs and phosphorylated Rho kinase. (**, p < 0.01 in Student’s t test; n = 6 blots from six different sets of littermates). (G) Tissue lysates from 14-d-old Rnd2 transgenic (Rnd2^Tg) and control mice were used for immunoblotting with the respective antibodies for pMbs, Mbs, pRho kinase, and Rho kinase. (H, I) Statistical data of phosphorylated Mbs and phosphorylated Rho kinase are shown (**, p < 0.01, *, p < 0.05 in Student’s t test; n = 3 blots from three different sets of littermates). (J) Tissue lysates from 28-d-old Rnd2 transgenic and control mice were used for immunoblotting with the respective antibodies for pMbs, Mbs, pRho kinase, and Rho kinase. (K, L) Statistical data of phosphorylated Mbs and phosphorylated Rho kinase are shown. (**, p < 0.01 in Student’s t test; n = 3 blots from three different sets of littermates).

FIGURE 9:

Plp1-Cre^ERT–mediated Rnd2 knockout mice exhibit decreased myelin thickness in adult mice at 2 wk after tamoxifen administration. (A) Eight- or nine-week-old mice were given a daily injection of 1 mg (100 μl) tamoxifen (TAM) for 5 consecutive days. Two weeks following the final TAM injection, antibodies against MBP (red) and Olig2 (green) were used for costaining in TAM-treated Rnd2^fl/fl; Plp1-Cre^ERTor control mouse brain cross-sections. The scale bars indicate 200 μm. (B) The intensity of MBP staining per square millimeter was relatively semiquantified (**, p < 0.01 in Student’s t test; n = 6–7 sections from three different sets of littermates). (C) The Olig2⁺ cells per square millimeter were counted (n = 7 sections from three different sets of littermates). (D) The brain lysates from TAM-treated Plp1-Cre^ERT–driven Rnd2 conditional knockout or control mice were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (E) CNPase expression levels are shown statistically (*, p < 0.05 in Student’s t test; n = 3 blots from three different sets of littermates). (F) Electron microscopic data of the corpus callosum of conditional knockout and control mice at 2 wk following TAM injection are shown. The scale bars indicate 2 μm. (G) Graph of g-ratios of myelinated axons for axon diameters is shown (n = 448 axons for three different knockout mice and n = 469 axons for their littermate controls). (H) Average g-ratios are shown (*, p < 0.05 in Student’s t test). (I) Graph of the distributions of g-ratios is shown.

FIGURE 10:

Plp1-Cre^ERT–mediated Rnd2 knockout mice exhibit increased myelin thickness in adult mice at 4 wk after tamoxifen administration. (A) Eight- or nine-week-old mice were given TAM for 5 consecutive days. Four weeks following the final TAM injection, antibodies against MBP (red) and Olig2 (green) were used for costaining in TAM-treated Rnd2^fl/fl; Plp1-Cre^ERTor control mouse brain cross-sections. The scale bars indicate 200 μm. (B) The intensity of MBP staining per square millimeter was relatively semiquantified (*, p < 0.05 in Student’s t test; n = 6 sections from three different sets of littermates). (C) The Olig2⁺ cells per square millimeter were counted (n = 6 sections from three different sets of littermates). (D) The brain lysates from TAM-treated Plp1-Cre^ERT–driven Rnd2 conditional knockout or control mice were immunoblotted with an antibody against MBP, CNPase, Rnd2, or actin. (E) CNPase expression levels are shown statistically (**, p < 0.01 in Student’s t test; n = 4 blots from three different sets of littermates). (F) Electron microscopic data of the corpus callosum of conditional knockout and control mice at 4 wk following TAM injection are shown. The scale bars indicate 2 μm. (G) Graph of g-ratios of myelinated axons for axon diameters is shown (n = 353 axons for three different knockout mice and n = 403 axons for their littermate controls). (H) Average g-ratios are shown (**, p < 0.01 in Student’s t test). (I) Graph of the distributions of g-ratios is shown.

FIGURE 11:

Schematic diagram of the role of Rnd2 in myelination. In earlier myelinating periods, Rnd2 down-regulates Rho kinase to promote myelination. In later myelinating periods, Rnd2 up-regulates Rho kinase to inhibit myelination.