SEARCH

SEARCH BY CITATION

Keywords:

  • gene expression;
  • oligodendrocyte;
  • spinal cord;
  • cerebellum;
  • corpus callosum

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Oligodendrocytes form myelin sheaths around axons in the central nervous system. Although cholesterol is one of the major lipid components in myelin sheath, the source of cholesterol for myelination remains to be defined. In this study, we report that low-density lipoprotein receptor (LDLR) and very low-density lipoprotein receptor (VLDLR) are selectively expressed in mature myelinating oligodendrocytes in the postnatal CNS. Both receptors are specifically expressed in differentiated oligodendrocytes in P7 spinal cord, but progressively down-regulated after P15. In adult animals, only LDLR expression can be detected in a small number of oligodendrocytes throughout the entire spinal cord. In the brain region, LDLR is expressed by the white matter oligodendrocytes of both cerebellum and cerebral cortex, whereas VLDLR has a weak expression in cerebellar oligodendrocytes. Together, our expression studies suggest that cholesterol uptake by LDLR and VLDLR may play an important role in the formation of myelin sheath. Developmental Dynamics 236:2708–2712, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Oligodendrocytes are myelin-forming glial cells in the central nervous system (CNS). The major function of myelin sheaths is to ensure insulation for rapid saltatory conduction of electrical signals along axons. The insulating properties of myelin sheath are largely due to its compact and multilayered stack of membranes and its richness in lipids. Myelin dry weight consists of 70% lipids and 30% proteins, and this lipid-to-protein ratio is generally the reverse in other cellular membranes (Baumann and Pham-Dinh,2001). In myelin, cholesterol constitutes nearly 40% of the total lipid content, as compared to less than 20% in other plasma membranes (Baumann and Pham-Dinh,2001; Morell and Jurevics,1996). Previous studies have suggested that cholesterol plays critical functions in the development of membrane properties. It can regulate membrane thickness and fluidity (Ohvo-Rekila et al.,2002), and limit ion leakage through membranes, both of which are important for the insulation properties of myelin (Haines,2001). Moreover, cholesterol, glycosphingolipids, and certain myelin proteins can form detergent-insoluble membrane complexes or membrane rafts (Brown and London,2000). The myelin membrane rafts can serve as platforms for protein sorting (Brown and London,2000) and signal transduction (Simons and Toomre,2000).

Cholesterol can be synthesized de novo by neurons and glial cells through a squalene synthase (SQS)-dependent pathway. Conditional ablation of the SQS gene (Fdft1) in oligodendrocytes resulted in a severely reduced rate of mylination in the white matter of the CNS, indicating that the endogenous cholesterol synthesis plays an important role in myelin formation (Saher et al.,2005). Similarly, several genetic disorders of the cholesterol biosynthetic pathway are associated with myelination defects in human (Bjorkhem and Meaney,2004; Wechsler et al.,2003). Intriguingly, hypomyelination caused by the SQS mutation can be overcome with time in older animals, suggesting that there might be a horizontal cholesterol transfer between oligodendrocytes and the surrounding cells or circulation. In support of this idea, low-density lipoprotein receptor related protein (LRP), one of the major receptors for cholesterol-loaded lipoprotein in the CNS (Herz and Bock,2002), was found to be expressed in oligodendrocytes (Saher et al.,2005). In the absence of cholesterol synthesis in the Fdft1 conditional mutants, LRP expression was slightly up-regulated in oligodendrocytes cells (Saher et al.,2005). In addition, the related member of the low-density lipoprotein receptor family, LRP2/Megalin, is also expressed in the white matter oligodendrocytes in postnatal mouse spinal cords (Wicher et al.,2006). LRP2 is a multifunctional receptor involved in cellular signaling and receptor-mediated endocytosis.

In this study, we examined the potential involvement of the low-density lipoprotein receptor (LDLR) and very low-density lipoprotein receptor (VLDLR) in cholesterol transport in myelinating oligodendrocytes. It was found that both receptors were transiently expressed in mature myelinating oligodendrocytes in the postnatal spinal cord. Expression of LDLR in oligodendrocytes was stronger and longer-lasting than that of VLDLR. In the rostral brain regions, only LDLR was expressed in the white matter oligodendrocytes, suggesting a regional difference in the involvement of these two receptor molecules in oligodendrocyte myelination.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

One major lipid component of the myelin sheath is cholesterol, which could be either synthesized de novo in oligodendrocyte cells or up-taken from the surrounding environment by receptor-mediated endocytosis. To determine whether the LDLR and VLDLR receptors participate in the cholesterol endocytosis during the myelin formation process, we carried out RNA in situ hybridization (ISH) with these two receptor molecules in mouse spinal cord tissues prepared from various postnatal stages. At the onset of myelination (P0), expression of both LDLR and VLDLR is detected in a few cells in the ventral white matter (Figs. 1A, 1A′, 2A, 2A′). As the myelination process proceeds, the number of LDLR+ and VLDLR+ cells in the white matter is significantly increased (Figs. 1B, 1B′, 2B, 2B′) and has reached the maximum at around P15 (Figs. 1C, 1C′, 2C, 2C′), the peak time of oligodendrocyte myelination in mouse spinal cords (Baumann and Pham-Dinh,2001). At this stage, many LDLR+ and VLDLR+ are also observed in the gray matter. At P30, LDLR expression is decreased but persists even in P60 adult spinal cords (Fig. 1D,E, 1D′,E′, 1F). In contrast, VLDLR expression cannot be observed in P30 and P60 (Fig. 2D, 2D′, 2E; data not shown), indicating the rapid down-regulation of VLDLR expression in the white matter glial cells.

thumbnail image

Figure 1. LDLR expression in the postnatal spinal cords. A–E: Spinal cord sections from P0, P7, P15, P30, and P60 were subject to ISH with LDLR riboprobe. A′–E′: Higher magnifications of A–E in the ventrolateral region. LDLR expression started to be detected at P0 (represented by an arrow) and persisted in all postnatal stages. F: Quantitative analyses of LDLR+ cells at various postnatal stages (n = 3).

Download figure to PowerPoint

thumbnail image

Figure 2. VLDLR expression in the postnatal spinal cords. A–D: Spinal cord sections from P0, P7, P15, P30 were subjected to ISH with VLDLR riboprobe. A′–D′: Higher magnifications of A–D in the ventrolateral region. VLDLR was weakly expressed in the spinal cord from P0 to P15, but down-regulated in P30. Weak expression in the white matter was represented by arrows. E: Quantitative analyses of VLDLR+ cells at various postnatal stages (n = 3).

Download figure to PowerPoint

The spatiotemporal pattern of LDLR and VLDLR expression suggests that these two receptor molecules are likely to be expressed in the mature myelinating oligodendrocytes. To test this possibility, we performed ISH with these two molecules on P7 spinal cord section followed by anti-Olig2, anti-APC (CC1) or anti-Nkx6.2 immunohistochemcal staining. Recent studies demonstrated that Olig2 is expressed in both oligodendrocyte progenitor cells and mature oligodendrocytes in postnatal spinal cord (Lu et al.,2000; Takeyabashi et al.,2000; Zhou et al.,2000), whereas APC and Nkx6.2 specifically label mature oligodendrocytes (Bhat et al.,1996; Southwood et al.,2004). Double labeling experiments revealed that at P7, the vast majority of LDLR+ and VLDLR+ cells co-expressed Olig2, APC, Nkx6.2, but not the astrocyte marker GFAP (Fig. 3A–C, G–I; data not shown). At P15, many LDLR+ cells in the gray matter also co-expressed Olig2 (Fig. 3D) and APC (Fig. 3E), but not the neuronal marker NeuN (Fig. 3F). Together, these results indicated that LDLR and VLDLR are indeed expressed by differentiated oligodendrocyte cells. Consistently, many Olig2+ cells did not co-express LDLR/VLDLR in P7 spinal cords (Fig. 3A,G,H), as Olig2 labels both OPCs and mature oligodendrocytes.

thumbnail image

Figure 3. Co-expression of LDLR and VLDLR with oligodendrocyte markers. P7 (A–C, G–I) and P15 (D–F) spinal cord sections were subjected to ISH (in blue) followed by immunohistochemical staining (in brown) with anti-Olig2 (A, D, G, H), anti-APC (B, E), anti-Nkx6.2 (C, I), or neuronal marker NeuN (F). Representative double positive oligodendrocytes in the ventrolateral white matter (A–C, G), dorsal white matter (H, I), or gray matter (D–F) are indicated by arrows. The Olig2+/LRLR- and Olig2+/VLDLR- cells in A, G, and H likely represent the immature OPC cells.

Download figure to PowerPoint

To further confirm that LDLR and VLDLR are selectively expressed in differentiated oligodendrocytes, we examined the expression of these two receptors in the Olig1 and Nkx2.2 mutant spinal cords by ISH. It was previously demonstrated that the differentiation of oligodendrocytes is delayed and/or reduced in the Olig1 and Nkx2.2 null mutant spinal cords (Qi et al.,2001; Lu et al.,2002). Consistent with the idea that LDLR and VLDLR are expressed in mature oligodendrocytes, expression of both genes was significantly decreased in the P3 Olig1 mutants and absent in Nkx2.2 mutant spinal cords (Fig. 4).

thumbnail image

Figure 4. Reduced expression of LDLR and VLDLR in Olig1 and Nkx2.2 mutant spinal cords. Spinal cord sections from P3 wild-type (A, D), Olig1−/− (B, E), and Nkx2.2−/− (C, F) spinal cord tissues were subjected to LDLR and VLDLR in situ hybridization. Only the ventrolateral regions of the spinal cord are shown.

Download figure to PowerPoint

Previous studies demonstrated that the related family member LRP2/Megalin is expressed in mature oligodendrocytes in the spinal cord, but not in the brain regions (Wicher et al.,2006). To examine whether a similar regional difference exists for LDLR and VLDLR expression, we carried out ISH of these two genes on early postnatal brain tissues. At P7, LDLR was clearly expressed in the white matter of the cerebellum (inset in Fig. 5A) and weakly expressed in the external granule layer. In contrast, VLDLR was highly expressed in the Purkinje cell layer and the underlying granule layer but weakly in the white matter oligodendrocytes (inset in Fig. 5B). Since oligodendrocyte differentiation occurs relatively late in the forebrain, we examined their expression in P15 cortical tissues. At this stage, expression of LDLR, but not VLDLR, was detected in the mature oligodendrocytes in the white matter (corpus callosum) of the cerebral cortex (Fig. 5C,D). Thus, LDLR and VLDLR appear to be differentially expressed in the rostral regions of the CNS.

thumbnail image

Figure 5. Expression of LDLR and VLDLR in the cerebellum and cerebral cortex. P7 cerebellum (A,B) and P15 cerebral cortex (C,D) were hybridized with LDLR (A, C) and VLDLR (B, D) riboprobes. Insets in A and B are higher magnifications showing the expression of LDLR and VLDLR by the white matter oligodendrocytes in the cerebellum. Expression of LDLR in the white matter oligodendrocytes of the cerebral cortex is indicated by arrows in C.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

In this study, we report the detailed analyses of LDLR and VLDLR expression in the developing mouse spinal cord by in situ hybridization. While expression of LDLR can be detected in myelinating oligodendroglial cells in all regions of the CNS, VLDLR is only weakly transcribed in the spinal cord and cerebellar regions. The lack of VLDLR expression in the forebrain region suggests the regional heterogeneity of oligodendrocytes in gene expression and possibly the utilization of exogenous cholesterol during the formation of myelin sheath by oligodendrocytes. In keeping with this notion, it has been recently demonstrated that LRP2/Megalin is similarly expressed in myelinated oligodendrocytes in the postnatal spinal cord, but not in the rostral brain regions (Wicher et al.,2006). LRP2/Megalin is a multifunctional member of the low-density lipoprotein receptor family involved in cellular signaling and receptor-mediated endocytosis of lipoproteins (Spoelgen et al.,2005). In addition, it was suggested the LRP molecule, another member of the low-density lipoprotein receptor family, is also expressed by spinal cord oligodendrocytes (Sahr et al.,2005). Since LDLR, VLDLR, LRP, and LRP2/Megalin are all involved in receptor-mediated endocytosis of cholesterol-containing lipoprotein complexes, the simultaneous expression of all these four structurally and functionally related receptor molecules in myelinating spinal cord oligodendrocytes strongly suggests that cholesterol uptake from surrounding environment plays an important role during the myelination process in the spinal cord, and possibly in the brain as well. Consistent with this concept, inhibition of the endogenous cholesterol synthesis by the disruption of SQS gene in oligodendrocytes only resulted in the delay of myelin sheath formation, and mutant animals can overcome hypomyelination at later stages probably by cholesterol uptake (Sahr et al.,2005). Since cholesterol is not imported to the brain from the circulation (Morell and Jurevics,1996), it is likely that in the conditional knockouts, cholesterol is horizontally transferred to mutant oligodendrocytes from adjacent wild-type cells such as neurons and/or astrocytes. Astrocytes are known to be a source of cholesterol for cultured neurons (Herz and Bock,2002; Pfrieger,2003), but whether astrocytes can deliver the cholesterol in vivo to neurons or oligodendrocytes remains to be determined.

The VLDLR, LDLR, LRP are the major apolipoprotein E-recognizing endocytic receptors involved in the clearance of triglyceride (TG)-rich lipoproteins from plasma (van Vlijmen et al.,1999). LRP2/Megalin also actively participates in the endocytosis of lipoproteins (May et al.,2005). Thus, it is possible that expression of these receptor molecules by differentiated oligodendrocytes might also be related to the uptake of other lipid components such as triglyceride for rapid membrane biogenesis during the myelination process. In addition, VLDLR and, to a lesser extent, LDLR are known to bind to reelin signal, which regulates neuronal migration and positioning during brain development (D'Arcangelo et al.,1999; Hiesberger et al.,1999; Trommsdorff et al.,1999), raising the other possibility that LDLR and VLDLR could function as reelin receptors during oligodendrocyte migration and positioning along axon fibers.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

Genotyping of Olig1 and Nkx2.2 Mutant Mice

The Olig1 and Nkx2.2 homozygous null pups were obtained by the interbreeding of heterozygous animals. Genomic DNA extracted from tails was used for genotyping by Southern analysis or by PCR. Genotyping of Olig1 and Nkx2.2 loci was described earlier (Lu et al.,2002; Qi et al.,2001).

In Situ RNA Hybridization and Double Labeling Experiments

Mouse spinal cord and brain tissues from various postnatal stages were fixed in 4% paraformaldehyde at 4°C overnight. Following fixation, tissues were transferred to 20% sucrose in PBS overnight, embedded in OCT media, and then sectioned (16 μm thickness) on a cryostat. Tissue sections from the wild-type and mutant animals were then subject to in situ hybridization (ISH) with LDLR (GenBank accession no. BC053041) and VLDLR (GenBank accession no. BC013622) riboprobes. ISH was performed as described in Schaeren-Wiemers and Gerfin-Moser (1993) with minor modifications. For double labeling experiments, tissues were first subject to LDLR and VLDLR ISH, followed by anti-Olig2 indirect immunohistochemical staining with ABC kit. Briefly, following ISH, slides were washed with PBS, blocked with PBS + 0.1% Triton + 1% BSA for 1 hr, and incubated with primary antibodies overnight at 4°C. Slides were then washed several times with PBST, and then incubate with secondary antibody for 1 hr at RT. After several washes with PBST, slides were detected with ABC kit (Vector Laboratories, Inc.) and 3, 3′-diaminobenzidine (DAB) solution following the manufacturer's instructions. Antibodies were used as follows: rabbit anti-Olig2 (a gift of Charles Stiles) at 1:2,000; mouse anti-APC (Ab-7, Oncogene Inc.) at 1:1,000; guinea pig anti-Nkx6.2 (a gift of Johan Ericson) at 1:6,000 and mouse anti-NeuN (Chemicon Inc.) at 1:1,000.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES

We thank Drs. Charles Stiles, David Rowitch, and Richard Lu for generously providing Olig1 mutant mice and the anti-Olig2 antibody, and Dr. Johan Ericson for the anti-Nkx6.2 antibody.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES