Platelet-derived growth factor promotes repair of chronically demyelinated white matter

Abstract

In multiple sclerosis, remyelination becomes limited after repeated or prolonged episodes of demyelination. To test the effect of platelet-derived growth factor-A (PDGF-A) in recovery from chronic demyelination we induced corpus callosum demyelination using cuprizone treatment in hPDGF-A transgenic (tg) mice with the human PDGF-A gene under control of an astrocyte-specific promoter. After chronic demyelination and removal of cuprizone from the diet, remyelination and oligodendrocyte density improved significantly in hPDGF-A tg mice compared with wild-type mice. In hPDGF-A tg mice, oligodendrocyte progenitor density and proliferation values were increased in the corpus callosum during acute demyelination but not during chronic demyelination or the subsequent recovery period, compared with hPDGF-A tg mice without cuprizone or to treatment-matched wild-type mice. Proliferation within the subventricular zone and subcallosal zone was elevated throughout cuprizone treatment but was not different between hPDGF-A tg and wild-type mice. Importantly, hPDGF-A tg mice had reduced apoptosis in the corpus callosum during the recovery period after chronic demyelination. Therefore, PDGF-A may support oligodendrocyte generation and survival to promote remyelination of chronic lesions. Furthermore, preventing oligodendrocyte apoptosis may be important not only during active demyelination but also for supporting the generation of new oligodendrocytes to remyelinate chronic lesions.

FIGURE 1

Detection of human platelet-derived growth factor-A (hPDGF-A) transgene expression in cuprizone-demyelinated corpus callosum. (A–F) In situ hybridization to detect hPDGF-A transgene mRNA transcripts in hPDGF-A transgenic (tg) mice (A–E) and wild-type mice (F). (E) Quantification of cells expressing hPDGF-A transgene mRNA transcripts in corpus callosum sections of hPDGF-A tg mice ± cuprizone (cup) for the time period indicated. (G–K) Immunofluorescence detection of the myc epitope tag of hPDGF-A transgene fusion protein in wild-type (G) and hPDGF-A tg mice (H–K). All panels are coronal brain sections aligned as in F with the corpus callosum midline along the left border (indicated by double-headed arrow) and showing the corpus callosum (cc) laterally to below the cingulum (cg). Mice were fed normal chow (A, H) or chow with 0.2% cuprizone with analysis after acute treatment (B, F, G, I; 6 wk cup), chronic treatment (C, J; 12 wk cup), or chronic treatment followed by a 6-week recovery period with consumption of normal chow (D, K; 12 wk cup 6 off). Scale bar (A–F as shown in D; G–K as shown in K) = 250 μm.

FIGURE 2

After chronic demyelination, spontaneous remyelination is increased in human platelet-derived growth factor-A (hPDGF-A) transgenic (tg) mice compared with wild-type mice. (A, B) Quantification of corpus callosum myelination estimated by immunofluorescence for myelin oligodendrocyte glycoprotein (MOG) in wild-type mice (A) and hPDGF-A tg mice (B). Black bars indicate perfusion after 3, 6, 9, or 12 weeks of continuous 0.2% cuprizone (cup) feeding, or 12 weeks of cuprizone followed by 6 weeks of consumption of normal chow. White bars indicate no cuprizone treatment. Pixel intensity values were normalized between tissue sections by thresholding to exclude values below the level of immunoreactivity in the dorsal fornix, which is not demyelinated by cuprizone. In mice of both genotypes, significant demyelination occurs with cuprizone treatment (*, p < 0.05 compared with no cup; 1-way analysis of variance [ANOVA] within each genotype). The hPDGF-A tg mice show significant remyelination of the corpus callosum during the recovery period with consumption of normal chow relative to the 12-week cuprizone chronic treatment (#, p < 0.05 for hPDGF-A tg 12 wk cup 6 off compared with hPDGF-A tg 12 wk cup; 1-way ANOVA within genotype). Further, myelination in the hPDGF-A tg mice is significantly increased from the wild-type mice only during the recovery after chronic cuprizone treatment (+, p < 0.05 for hPDGF-A tg 12 wk cup 6 weeks off compared with hPDGF-A wild-type 12 wk cup 6 off; 2-way ANOVA for genotype and treatment). With this pattern during the recovery period, the values for hPDGF-A tg mice have returned to within normal levels (p > 0.05 for 12 wk cup 6 off compared with no cup; 1-way ANOVA within genotype), whereas wild-type values remain significantly below normal levels (p < 0.001 for 12 wk cup 6 off compared with no cup; 1-way ANOVA within genotype). All statistical analyses are based on a sample size of 3 to 5 mice per condition. (C–H) Representative images of MOG immunostaining of coronal sections through the corpus callosum (areas shown as noted in Fig. 1; cc, corpus callosum; df, dorsal fornix; cg, cingulum, double-headed arrow along cc midline). (C, F) Wild-type and hPDGF-A tg mice, respectively, maintained on normal chow, without cuprizone treatment. (D, G) Wild-type and hPDGF-A tg mice, respectively, fed cuprizone continuously for 12 weeks. (E, H) Wild-type and hPDGF-A tg (H) mice, respectively, fed cuprizone continuously for 12 weeks, followed by 6 weeks of recovery. Scale bar (C–H shown in H) = 250 μm.

FIGURE 3

Human platelet-derived growth factor-A (hPDGF-A) transgenic (tg) mice have increased oligodendrocyte progenitor (OP) amplification in the corpus callosum only during the acute stage of cuprizone (cup) demyelination. (A–D) Quantification of the OP response in the corpus callosum of wild-type (A, C) and hPDGF-A tg (B, D) mice. Black bars indicate perfusion after 3, 6, 9, or 12 weeks of continuous 0.2% cuprizone feeding or 12 weeks of cuprizone followed by 6 weeks of normal chow. White bars indicate no cuprizone treatment. (A, B) Density of the total OP population, identified by PDGFαR mRNA expression. (C, D) Density of proliferating OP cells identified by PDGFαR mRNA and incorporation of BrdU during a 4-hour terminal pulse. Before the start of cuprizone treatment, hPDGF-A mice of both genotypes have similar densities of total OP cells (A, B) and similar extents of proliferation among the OP pool (C, D). After 6 weeks of cuprizone treatment in hPDGF-A tg mice, significant increases are observed in the total OP cell density (B, *p < 0.001) and the extent of OP proliferation (D, *p < 0.001) compared with hPDGF-A tg mice without cuprizone treatment. All statistical analyses are based on a sample size of 3 to 5 mice per condition. (E–J) Representative coronal sections showing PDGFαR mRNA in situ hybridization (blue/black cytoplasm) and BrdU incorporation (brown nuclei) in the corpus callosum (E: cc, corpus callosum; df, dorsal fornix, arrow along cc midline). (E, F) Wild-type and hPDGF-A tg mice, respectively, maintained on normal chow without cuprizone treatment. (G, H) Wild-type and hPDGF-A tg mice, respectively, fed cuprizone continuously for 6 weeks. (I, J) Wild-type and hPDGF-A tg mice, respectively, fed cuprizone continuously for 12 weeks followed by a 6-week recovery period of normal chow. Scale bar ((E–J) shown in H) = 250 μm.

FIGURE 4

Subventricular zone (SVZ) activation continues throughout acute and chronic demyelination. (A–C) Representative coronal sections through the corpus callosum (cc, double arrow along midline) and SVZ showing platelet-derived growth factor-α receptor (PDGFαR) mRNA in situ hybridization (blue/black cytoplasm) to identify oligodendrocyte progenitor (OP) cells and bromodeoxyuridine (BrdU) incorporation (brown nuclei) during a 4-hour terminal pulse. (A) Within the SVZ, BrdU labeling is evident in nontreated adult (8 weeks of age) human platelet-derived growth factor-A (hPDGF-A) transgenic (tg) (shown) and wild-type mice (not shown). (B) After 5 weeks of cuprizone (cup), hPDGF-A tg mice demonstrate robust proliferation and OP amplification in the lesioned area of the corpus callosum relative to the extent of response in the SVZ. (C) After cuprizone for 12 weeks with a 6-week recovery period, proliferation is marked within the SVZ relative to the attenuated response in the corpus callosum in both wild-type (shown) and hPDGF-A tg mice (not shown). Scale bars (D–I) = 200 μm. Quantification within germinal zones of overall proliferative index (BrdU labeling in D, G), total OP density (E, H), and density of proliferating OP cells (F, I) in wild-type (D–F) and hPDGF-A tg (G–I) mice. White bars indicate no cuprizone treatment. Gray bars indicate acute cuprizone treatment (3 or 6 weeks of cuprizone or 6 weeks of cuprizone with 3 weeks for recovery). Black bars indicate 12 weeks of continuous cuprizone feeding (12 weeks of cuprizone or 12 weeks of cuprizone with 3 or 6 weeks for recovery). *, p < 0.05 for 1-way analysis of variance within each genotype.

FIGURE 5

Oligodendrocyte repopulation of chronic lesions is enhanced in human platelet-derived growth factor-A (hPDGF-A) transgenic (tg) mice. (A, B) Quantification of the density of oligodendrocytes, identified by in situ hybridization for PLP mRNA, in the corpus callosum (cc) of wild-type (A) and hPDGF-A tg (B) mice. Black bars indicate perfusion after 3, 6, 9, or 12 weeks of continuous 0.2% cuprizone (cup) feeding or 12 weeks of cuprizone followed by 6 weeks of normal chow. White bars indicate no cuprizone treatment. At least 3 sections were quantified per mouse from at least 3 mice per condition. Before the start of cuprizone treatment, hPDGF-A tg mice have a higher density of oligodendrocytes than wild-type mice at 8 weeks of age (ˆ, p = 0.0009; Student's t-test between genotypes of mice without cuprizone treatment). During cuprizone treatment of both wild-type and hPDGF-A tg mice, the density of oligodendrocytes is significantly decreased from normal levels (*, p < 0.0001 compared with no cup; 1-way analysis of variance (ANOVA) within each genotype). Mice of both genotypes show significant oligodendroglial repopulation of the corpus callosum during the recovery period (#, p < 0.05 for 12 wk cup 6 off compared with 12 wk cup; 1-way ANOVA within genotype). However, the oligodendroglial repopulation in the hPDGF-A tg mice is significantly increased from that in the wild-type mice (+, p < 0.05 for hPDGF-A tg 12 wk cup 6 off compared with wild-type 12 wk cup 6 off; 2-way ANOVA for genotype and treatment). With this pattern during the recovery period, the values for hPDGF-A tg mice have returned to within normal levels (p > 0.05 for 12 wk cup 6 off compared with no cup; 1-way ANOVA within genotype), whereas wild-type values remain significantly below normal levels (*, p < 0.05 for 12 wks cup 6 off compared with no cup; 1-way ANOVA within genotype). All statistical analyses are based on a sample size of 3 to 5 mice per condition. (C–H) Representative coronal sections showing PLP mRNA in situ hybridization in the corpus callosum (C: cc, corpus callosum; df, dorsal fornix, double headed arrow along cc midline). (C, F) Wild-type and hPDGF-A tg mice, respectively, maintained on normal chow, without cuprizone treatment. (D, G) Wild-type and hPDGF-A tg mice, respectively, fed cuprizone continuously for 12 weeks. (E, H) Wild-type and hPDGF-A tg mice, respectively, fed cuprizone continuously for 12 weeks, followed by 6 weeks of recovery. Scale bar (C–H shown in H) = 250 μm.

FIGURE 6

During recovery from chronic demyelination, apoptosis is reduced in human platelet-derived growth factor-A (hPDGF-A) transgenic (tg) mice. (A, B) Quantification of the density of apoptotic cells, identified by a modified terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay, in the corpus callosum of wild-type (A) and hPDGF-A tg mice (B). White bars indicate no cuprizone (cup) treatment. Gray bars indicate perfusion after 3 or 6 weeks of cuprizone treatment. Black bars indicate perfusion after 12 weeks of continuous cuprizone feeding or 12 weeks of cuprizone followed by 6 weeks of normal chow. During cuprizone treatment of both wild-type and hPDGF-A tg mice, the density of apoptotic cells is significantly increased from normal levels (*, p < 0.05 compared with no cup; 1-way analysis of variance (ANOVA) within each genotype). With 6 weeks of cuprizone treatment, TUNEL labeling in the hPDGF-A tg mice is significantly increased from that in the wild-type mice (#, p < 0.05 for hPDGF-A tg 6 wk cup compared with wild-type 6 wk cup; 2-way ANOVA for genotype and treatment). Elevated apoptosis values continued after removal of cuprizone from the diet of wild-type mice (*, p < 0.05 compared with no cup). Although apoptosis remains elevated somewhat during recovery in hPDGF-A tg mice, the values are no longer significantly above those in nontreated hPDGF-A tg mice. Further, apoptosis in the hPDGF-A tg mice is significantly decreased from that in the wild-type mice (+, p < 0.05 for hPDGF-A tg 12 wk cup 6 wk off compared with wild-type 12 wk cup 6 wk off; 2-way ANOVA for genotype and treatment). All statistical analyses are based on a sample size of 3 to 5 mice per condition. (C) Representative coronal section showing the corpus callosum (cc, double headed arrow along midline; cg, cingulum) in a wild-type mouse treated with cuprizone for 12 weeks followed by a 6-week period for recovery. TUNEL (brown) signal is present in individual cells distributed throughout the corpus callosum and in groups of cells aligned as is characteristic of interfascicular oligodendrocytes (single arrows and inset). Nuclei stained with methyl green. Scale bars = 250 and 10 μm. (D–F) Immunostaining for Olig2 (D, and red in F) to identify oligodendrocyte lineage cells in combination with TUNEL (E, and green in F) within the corpus callosum in a wild-type mouse treated with cuprizone for 12 weeks followed by a 6-week period for recovery. Scale bar (D–F shown in F) = 10 μm. (G–H) Representative coronal sections showing PDGFαR mRNA in situ hybridization (blue/black cytoplasm) and TUNEL labeling (brown nuclei) in the corpus callosum (G: cc, corpus callosum, double arrow along cc midline). (G) After 6 weeks of cuprizone treatment, hPDGF-A tg mice show OP amplification (as in Fig. 3). TUNEL labeling is high but does not colocalize with PDGFαR expression, as expected for apoptosis of oligodendrocytes from ongoing cuprizone treatment. Scale bar = 200 μm. (H, I) During recovery after 12 weeks of cuprizone treatment, TUNEL and PDGFαR mRNA transcripts can be detected separately (H; wild-type mouse) but colocalization was also observed (I; wild-type mouse). Scale bars = (H, I) 10 μm.

FIGURE 7

Apoptosis continues during recovery from chronic cuprizone (cup) in C57BL/6 mice. Quantification of the density of apoptotic cells, identified by a modified TUNEL assay, in the corpus callosum of C57BL/6 mice. White bars indicate no cuprizone treatment. Gray bars indicate perfusion after acute cuprizone treatment periods of 3 or 6 weeks or 6 weeks of cuprizone followed by a recovery period of 3 or 6 weeks of normal chow. Black bars indicate perfusion after chronic demyelination with 12 weeks of continuous cuprizone feeding or 12 weeks of cuprizone followed by 3 or 6 weeks of normal chow. TUNEL labeling increases significantly by 3 weeks of cuprizone treatment (p < 0.05, compared with no cup; 1-way analysis of variance with n = 3 to 6 mice per time point). After acute demyelination, TUNEL levels decrease during the recovery period. In contrast, after chronic demyelination apoptosis continues at high levels during the subsequent recovery periods of 3 and 6 weeks.