Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2014 Jul 17;158(2):383-396.
doi: 10.1016/j.cell.2014.04.052. Epub 2014 Jul 10.

Oligodendrocyte-encoded HIF Function Couples Postnatal Myelination and White Matter Angiogenesis

Free PMC article

Oligodendrocyte-encoded HIF Function Couples Postnatal Myelination and White Matter Angiogenesis

Tracy J Yuen et al. Cell. .
Free PMC article


Myelin sheaths provide critical functional and trophic support for axons in white matter tracts of the brain. Oligodendrocyte precursor cells (OPCs) have extraordinary metabolic requirements during development as they differentiate to produce multiple myelin segments, implying that they must first secure adequate access to blood supply. However, mechanisms that coordinate myelination and angiogenesis are unclear. Here, we show that oxygen tension, mediated by OPC-encoded hypoxia-inducible factor (HIF) function, is an essential regulator of postnatal myelination. Constitutive HIF1/2α stabilization resulted in OPC maturation arrest through autocrine activation of canonical Wnt7a/7b. Surprisingly, such OPCs also show paracrine activity that induces excessive postnatal white matter angiogenesis in vivo and directly stimulates endothelial cell proliferation in vitro. Conversely, OPC-specific HIF1/2α loss of function leads to insufficient angiogenesis in corpus callosum and catastrophic axon loss. These findings indicate that OPC-intrinsic HIF signaling couples postnatal white matter angiogenesis, axon integrity, and the onset of myelination in mammalian forebrain.


Figure 1
Figure 1. Oligodendrocyte-specific VHL deletion inhibits differentiation and myelination
(A) Schematic of anatomical regions of corpus callosum (CC), cerebral cortex (CTX), and ventricle (V) presented in (B) and experimental timeline for chronic hypoxic rearing. (B) Images showing hypomyelination, OL-lineage HIF1α expression, and OPC maturation arrest in CC of hypoxic WT mice or normoxic Sox10-cre, VHL(fl/fl) mice at P11. Arrowheads denote double-positive cells. Scale bar: 100µm (MBP), 50µm (Olig2). (C) Immunopurified OPCs exposed to hypoxia or isolated from Plp-creERT2, VHL(fl/fl) mice show differentiation block. Scale bar: 100µm. (For quantifications, mean+SEM; n≥3 experiments/genotype; **p<0.01, ***p<0.001; one-way ANOVA with Dunnett’s multiple comparison test) See also Figure S1.
Figure 2
Figure 2. OPC-encoded HIF1/2α function mediates hypoxia-induced hypomyelination
(A) Schematic and timeline for cerebellar slice cultures (CSC) exposed to hypoxia. (B) Removing HIF1/2α function in OLs significantly reduces hypoxia-induced hypomyelination in CSC. Scale bars: 25µm (Caspr), 50µm (MBP/NFH). n≥6 experiments/condition. (C) Quantification of myelination in CSC. (D) Additional quantification of myelination in CSC. (E) Removing OPC HIF1/2α function significantly reduces hypoxia differentiation block in CSC. Scale bar: 100µm. (F) Quantification of OL differentiation showing Nkx2.2/Olig2 (OPCs) numbers decreased and CC1/Olig2 (mature OLs) numbers increased. Data were analyzed by t-test and significant differences (**p<0.01) are shown. (G) Model for HIF-induced OPC differentiation block. (For quantifications in C and D, mean+SEM; **p<0.01, ***p<0.001; one-way ANOVA with Dunnett’s multiple comparison test) See also Figure S2.
Figure 3
Figure 3. HIF stabilization in OPCs activates canonical Wnt signaling
(A) Scheme showing Wnt signaling and inhibition of ligand secretion and canonical activity by porcupine inhibitor IWP2 and XAV939, which stabilizes Axin2 to promote β-catenin degradation. (B) Western blots of P11 white matter demonstrating upregulation of activated β-catenin and Axin2 levels in WT mice reared in hypoxia and normoxic Sox10-cre, VHL(fl/fl) mice (n=3 animals/genotype). (C) IWP2 prevents hypomyelination in CSC exposed to hypoxia or from Plp-creERT2, VHL(fl/fl) mice. Scale bars: 25µm (Caspr), 50µm (MBP/NFH). (D) OPC maturation arrest in Plp-creERT2, VHL(fl/fl) OPCs is attenuated by IWP2 or XAV939. Scale bar: 100µm. (E) Immunopurified OPCs cultured in hypoxic conditions or exposed to DMOG specifically upregulate Wnt7a and Wnt7b as shown by qRT-PCR (n=3). (F) Positive control for ChIP analysis at Epo locus. (G) Mouse HIF1/2α knockout and control embryonic fibroblasts cultured with/without DMOG (16h) assayed by ChIP. Following immunoprecipitation with antibodies against HIF1α or control (mouse IgG), DNA extracts were assessed by qRT-PCR. HIF1α bound to the Wnt7a locus at one HRE, and Wnt7b locus via two HREs. Binding was not observed in DMOG-treated HIF1/2α mutant cells, or non-DMOG-treated controls. (H) Wnt7a proteins cause hypomyelination and OPC maturation arrest in CSC, which is reversed with XAV939. Scale bars: 25µm (Caspr), 50µm (MBP/NFH). (For quantification in C, D, and H mean+SEM; n≥3 experiments; *p<0.05, **p<0.01, ***p<0.001; one-way ANOVA with Dunnett’s multiple comparison test) See also Figure S3.
Figure 4
Figure 4. HIF stabilization in OPCs promotes angiogenesis in vivo
(A) Increased angiogenesis in Sox10-cre, VHL(fl/fl) mice as shown by expression of endothelial marker, CD31. Dense regions of Olig2 staining indicate white matter tracts in the corpus callosum. Scale bar: 100µm. n≥3 animals/genotype. (B) Quantification of endothelial/vessel area (CD31+) demonstrates significant increases in hypoxic WT and normoxic Sox10-cre, VHL(fl/fl) mice. Data were analyzed by one-way ANOVA with Dunnett’s multiple comparison test, and significant differences (*p<0.05, **p<0.05, ***p<0.001) are shown. (C) Endothelial marker Isolectin demonstrating increased angiogenesis in Sox10-cre, VHL(fl/fl) mice. Scale bar: 100µm. (D) Isolectin perfusion in WT and Sox10-cre, VHL(fl/fl) mice indicating perfusion of blood vessels. Scale bar: 100µm. (E) Increased Lef1 expression in endothelia of Sox10-cre, VHL(fl/fl) mice. Of these, the majority co-labeled with the proliferation marker, Ki67. Scale bar: 50µm. n≥3 animals/genotype. (F) Quantification of Ki67+/Lef1+ endothelial cells in corpus callosum and cortex. Data were analyzed by t-test and significant differences (***p<0.001) are shown. See also Figure S4, Table S1 and S3.
Figure 5
Figure 5. OPCs directly promote angiogenesis in a Wnt-dependent manner
(A) Scheme showing transwell co-culture assay for OPCs and bEND.3 cells. OPCs from Plp-creERT2, VHL(fl/fl) mice induce endothelial cell proliferation in a Wnt-dependent manner. Scale bar: 100µm. (B) Quantification of endothelial cell proliferation in transwell assay at 24h and 48h. (C) Wnt7a treatment of bEND.3 cells induces Lef1 expression (arrowheads). Scale bar: 30µm. (D) Transwell co-cultures of Plp-creERT2, VHL(fl/fl) OPCs and bEND.3 cells promotes endothelial cell tube formation in a Wnt-dependent manner. Scale bar: 500µm. (E) Schematic showing retina endothelial tip sprouting assay. Conditioned medium from Plp-creERT2, VHL(fl/fl) OPCs promoted endothelial tip sprouting and filopodia extension in a Wnt-dependent manner. Scale bar: 25µm. (For all quantifications mean+SEM; n≥3 experiments (A,B), n≥2 (D,E); **p<0.01, ***p<0.001; one-way ANOVA with Dunnett’s multiple comparison test) See also Figure S5.
Figure 6
Figure 6. Oligodendrocyte HIF1/2α function is required for postnatal angiogenesis and maintenance of white matter integrity
(A) DAPI stained sections at E18, P4, and P7 of mutant and control brains show white matter cysts (white asterisk) and dysgenesis by P7 in Olig1-cre, HIF1/2α(fl/fl) mice. n≥3 animals/genotype. Scale bar: 100µm. (B) Olig2+ cells are reduced by ~40% compared to WT at E18 and ~15% at P4. Vessel density in SVZ at E18 is similar in WT and mutant mice, whereas Olig1-cre, HIF1/2α(fl/fl) mice show significantly decreased vessel density at P4 in corpus callosum. Data were analyzed by t-test and significant differences (*p<0.05, **p<0.01) are shown. Scale bar: 100µm. (C) White matter cysts and increased apoptotic cells (Casp3+) in P4 in Olig1-cre, HIF1/2α(fl/fl) mice. Data were analyzed by t-test and the significant difference (*p<0.05) is shown. Scale bar: 100µm (merged), 50µm (Casp3) (D) Widespread axonal damage, indicated by SMI32+ staining, observed at P4 throughout the corpus callosum of Olig1-cre, HIF1/2α(fl/fl) mice. Scale bar: 100µm. (E) Robust Casp3 staining in axons of P4 Olig1-cre, HIF1/2α(fl/fl) corpus callosum. Note relative paucity of staining in cortex. Scale bar: 100µm. See also Figure S6, Table S2 and S3.
Figure 7
Figure 7. Loss of OPC HIF1/2 α function is permissive for cortical development and angiogenesis
(A) DAPI stain of primary motor cortex in WT versus Olig1-cre, HIF1/2α(fl/fl) mice at P7 showing a thinner cortex in Olig1-cre, HIF1/2α(fl/fl) with the cortical layers and overall structure intact. Cortical layers are labeled to the left, and the asterisk denotes white matter cyst. n=3 animals/genotype. Scale bar: 200µm. (B) OL numbers are reduced by approximately 23% in Olig1-cre, HIF1/2α(fl/fl) cortex, but vessel density (%CD31) is not statistically different. Data were analyzed by a two-tailed Student’s t-test and the significant difference (*p<0.05) is shown. n=3 animals/genotype. Scale bar: 100µm. (C) Images of NeuN (green, pan-neuron marker), SatB2 (red, layer 2/3 callosal projection neurons), and DAPI providing further evidence that the cortex is grossly intact with ample numbers of callosal projection neurons. Note in higher magnification panels (C’ and C”) that cell density is grossly normal in Olig1-cre, HIF1/2α(fl/fl) cortex. Scale bars: 200µm; 100µm (insets). (D) Images of the corpus callosum stained for BNIP3 (red) and Olig2 (green). In WT, BNIP3 is expressed in a subset of Olig2+ cells (arrows, D’ insets). In Olig1-cre, HIF1/2α(fl/fl) mice, while BNIP3 is not expressed in Olig2+ cells, aberrant expression of BNIP3 in non-Olig2+ cells (arrowheads, D” insets) is indicative of the general hypoxic microenvironment. Scale bars: 100µm; 20µm (insets). (E) Images of BNIP3 staining in dorsal cortex (top row) and ventral cortex (bottom row). BNIP3 is enriched in ventral, but not dorsal cortex, suggesting selective hypoxia in grey matter regions adjacent to the corpus callosum, but not more dorsal areas. n=3 animals/genotype. Scale bar: 100µm. See also Figure S7 and Table S7.

Comment in

Similar articles

See all similar articles

Cited by 97 articles

See all "Cited by" articles

Publication types

MeSH terms


LinkOut - more resources