SEARCH

SEARCH BY CITATION

Keywords:

  • Neural crest-derived cells;
  • NG2;
  • p75;
  • PDGFRβ;
  • Pericytes

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

Neural crest (NC) cells originate from the neural folds and migrate into the various embryonic regions where they differentiate into multiple cell types. A population of cephalic neural crest-derived cells (NCDCs) penetrates back into the developing forebrain to differentiate into microvascular pericytes, but little is known about when and how cephalic NCDCs invade the telencephalon and differentiate into pericytes. Using a transgenic mouse line in which NCDCs are genetically labeled with enhanced green fluorescent protein (EGFP), we observed that NCDCs started to invade the telencephalon together with endothelial cells from embryonic day (E) 9.5. A majority of NCDCs located in the telencephalon expressed pericyte markers, that is, PDGFRβ and NG2, and differentiated into pericytes around E11.5. Surprisingly, many of the NC-derived pericytes express p75, an undifferentiated NCDC marker at E11.5, as well as NCDCs in the mesenchyme. At the same time, a minor population of NCDCs that located separately from blood vessels in the telencephalon were NG2-negative and some of these NCDCs also expressed p75. Proliferation and differentiation of pericytes appeared to occur in a specific mesenchymal region where blood vessels penetrated into the telencephalon. These results indicate that (i) NCDCs penetrate back into the telencephalon in parallel with angiogenesis, (ii) many NC-derived pericytes may be still in pre-mature states even though after differentiation into pericytes in the early developing stages, (iii) a small minority of NCDCs may retain undifferentiated states in the developing telencephalon, and (iv) a majority of NCDCs proliferate and differentiate into pericytes in the mesenchyme around the telencephalon.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

The neural crest (NC) is a transient embryonic tissue that originates at the neural folds during early vertebrate development. Cells delaminate from the NC migrate into various locations, and these neural crest-derived cells (NCDCs) differentiate into multiple cell types, including melanocytes of the skin and the neurons and glia of the autonomic and enteric nervous systems (Le Douarin & Kalcheim 1999; Sieber-Blum 2012). In addition to the abovementioned cell types, NCDCs in the craniofacial region give rise to the meninges (i.e. the dura mater, pia mater, and arachnoid mater), the bone and cartilage of the face, and the smooth muscles and pericytes of blood vessels (Sauka-Spengler & Bronner-Fraser 2008). Based on various derivatives generated by the NC, a major unresolved question is whether the NC is composed of homogeneous multipotent stem/precursor cells or, alternatively, of a mixed heterogeneous population of cells, including multipotent stem/precursor cells, fate-restricted cells, and cells that are committed to a particular cell lineage (Le Douarin et al. 2008). Previous in vitro studies in avian (Le Douarin & Kalcheim 1999) and in mammals (Stemple & Anderson 1992; Ito et al. 1993) have revealed that the developmental capacities of single NCDC are markedly broad and are heterogeneous. It is further revealed, from transplantation experiments using quail-chick chimeras, that the fate of NCDCs is not fixed in the neural primordium and rather determined, at least in part, by their local environment (Le Douarin & Kalcheim 1999). Therefore, the NC is a valuable model system for the study of cell lineage diversification that is controlled, at least in part, by environmental cues.

Previous studies with quail-chick chimeras have demonstrated that cephalic NCDCs re-invade the forebrain from embryonic day (E) 4 and differentiate into microvascular pericytes by E8, but not into endothelial cells (Etchevers et al. 2001; Korn et al. 2002). Etchevers et al. also indicate that quail NCDCs are first seen within the avian host telencephalon associated with the first capillaries on E4, although these capillaries have not been characterized by using endothelial cell markers. In mammals, a study using a NCDC-tracing Ht-PA-Cre/R26R mouse line has suggested that NCDCs associating with the vasculature of the telencephalon at E12.5 may be pericytes, although the character of these NCDCs has not been revealed by using pericyte markers (Pietri et al. 2003). Thus, penetration and differentiation of cephalic NCDCs that invade the telencephalon are still largely unknown.

Pericytes provide structural support for the vasculature throughout the body and control endothelial cell functions (Gerhardt & Betsholtz 2003; Armulik et al. 2005). It is known that PDGF-B/PDGFRβ signaling is essential for generation of pericytes and their recruitment to blood vessels (Hellström et al. 1999). Mice deficient either for PDGF-B or PDGFRβ lack brain pericytes and eventually die for hemorrhage and edema in late embryogenesis (Lindahl et al. 1997; Hellstrom et al. 2001, 1999). In the brain, pericytes are necessary for the formation and regulation of the blood–brain barrier (Armulik et al. 2010; Daneman et al. 2010) and control capillary diameter, probably in response to neural activity (Peppiatt et al. 2006). Although brain pericytes have these significant roles, little is known about the generation of NC-derived pericytes in the brain. To address these issues, we performed, by using genetically engineered reporter mice, detailed lineage tracing of cephalic NCDCs that penetrated into the telencephalon, and examined when and how these NCDCs invaded the telencephalon and differentiated into pericytes in the early developing stages.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

Animals

The myelin protein zero (P0)-Cre mouse line (Yamauchi et al. 1999) and CAG-CAT-EGFP-reporter mice (Kawamoto et al. 2000) were maintained on an ICR background and crossed to obtain the founder P0-Cre+/−/EGFP+/− mice (hereafter, P0-Cre/EGFP mice; Kanakubo et al. 2006). Then embryos were obtained by crossing P0-Cre/EGFP males with wild-type ICR females. Noon of the day when the vaginal plug was observed was designated as E0.5. To determine the genotypes of the examined mice, polymerase chain reaction (PCR) experiments were performed on genomic DNA as previously reported (Yamauchi et al. 1999; Kawamoto et al. 2000). The Mesp1-Cre (Saga et al. 1999) and R26R reporter mice (Soriano 1999) were maintained on a mixture of C57BL/6 and MCH backgrounds. The Mesp1-Cre/R26R mouse embryos were obtained by crossing Mesp1-Cre mice with homozygous R26R reporter mice. The following stages of P0-Cre/EGFP mice were examined: E9.5 (n = 28), E10.5 (n = 9), E11.0 (n = 6), E11.5 (n = 27). To examine usability of PDGFRβ and NG2 for pericyte markers, we used three E11.5 P0-Cre+/- embryos from littermates of P0-Cre/EGFP mice. Two E11.5 Mesp1-Cre/R26R embryos were examined. For the embryonic analyses, the gestational ages were determined by somite number as follows: E9.5 (21–29 somites), E10.5 (35–39 somites), E11.0 (40–44 somites), and E11.5 (45–47 somites). All of the animal experiments were approved by the Committee for Animal Experimentation of the Tohoku University Graduate School of Medicine and were performed in accordance with the guidelines for the Care and Use of Laboratory Animals published by the National Institutes of Health.

Immunohistochemistry

Immunohistochemistry on frozen sections was performed as previously described with minor modifications (Takahashi & Osumi 2002). The mouse embryos were fixed by immersion in 4% paraformaldehyde (PFA)/phosphate-buffered saline (PBS) overnight at 4°C, and immersed in 10% sucrose in PBS for 3 h to overnight and in 20% sucrose in PBS overnight at 4°C. The embryos were embedded in OCT (Sakura) and cut into 12, 30, or 40 μm coronal sections using a cryostat (CM-3050, Leica). The antibodies used in this study are summarized in Table 1. The primary and secondary antibodies were diluted with PBS containing 3% bovine serum albumin (BSA) (Sigma-Aldrich) and 0.3% Triton X-100 (Sigma-Aldrich). The sections were incubated with primary antibodies overnight at 4°C and with secondary antibodies for 1 h at room temperature. Double immunolabeling with primary antibodies from the same host species (rabbit anti-laminin and NG2 antibodies) was carried out by two-step protocol (Jackson ImmunoResearch Laboratories). First, after immunoreaction with rabbit anti-NG2 antibody, sections were incubated with Alexa Fluor 488-conjugated monovalent Fab fragment anti-rabbit IgG, and subsequently treated with unconjugated monovalent Fab fragment anti-rabbit IgG for 24 h at room temperature to block any nonspecific binding site. As the second staining, sections were incubated with rabbit anti-laminin antibody and followed with anti-rabbit Cy3-conjugated secondary antibody. The sections were subsequently counterstained with 4′, 6-diamidino-2-phenylindole (DAPI) (1:1000, Sigma-Aldrich). For the enzymatic detection of EGFP, the ABC kit (Vector Laboratories) and an enhanced diaminobenzidine (DAB) kit (Pierce) were used. Laminin was detected via immunofluorescence staining. For analyses of pericytes and endothelial cells at E11.5, frozen sections (40 μm) were immunostained using a floating method. The images were recorded using an Axioplan II fluorescent microscope that was equipped with an AxioCam CCD camera (Carl Zeiss), laser scanning confocal microscopy (LSM5 PASCAL or 510 META, Carl Zeiss), or a fluorescent microscope BZ-9000 (Keyence). Thick (1 mm) coronal slices were obtained using a vibratome (Microslicer, Dosaka). Immunostaining was performed on these slices as previously described (Osumi et al. 1997). The images were recorded using a cooled color CCD camera (Penguin 600CL, Pixera).

Table 1. List of antibodies
AntibodiesSpecies/classDilutionVendor (catalog No.)
Primary antibodies
BrdUMouse IgG11:50BD Bioscience (347580)
CreRabbit IgG1:5000Novagen (69050)
CreMouse IgG11:1000Millipore (MAB3120)
eNOSMouse IgG11:300BD Bioscience (610296)
Flk1Goat IgG1:500R&D Systems (AF644)
Flk1Rat IgG2a, κ1:500BD Bioscience (550549)
GFPChicken IgY1:1000Abcam (ab13970)
LacZChicken IgY1:500Abcam (ab9361)
LamininRabbit IgG1:4000Sigma-Aldrich (L9393)
NG2Rabbit IgG1:300Millipore (AB5320)
p75Mouse IgG2a1:2000Nobus Biologicals (M-009-100)
PDGFRβRabbit IgG1:200Cell Signaling (#3169)
PDGFRβRat IgG2a, κ1:200eBioscience (14-1402)
PECAM1Rat IgG2a, κ1:500BD Bioscience (550274)
Secondary antibodies
Chicken Alexa 488Goat1:400Molecular Probes (A11039)
Chicken DyLight 488Donkey1:400Jackson (703-485-155)
Goat IgG Cy5Donkey1:200Jackson (705-175-147)
Mouse IgG Cy3Donkey1:400Jackson (715-165-151)
Mouse IgG Cy5Donkey1:200Jackson (715-175-151)
Rabbit Fab fragment IgGDonkey1:10Jackson (711-007-003)
Rabbit Fab fragment IgG Alexa488Donkey1:500Jackson (711-547-003)
Rabbit IgG Cy3Donkey1:400Jackson (711-165-152)
Rabbit IgG-HRPGoat1:400Promega (W401B)
Rat DyLight 649Goat1:200Rockland (612-143-120)
Rat IgG Cy3Donkey1:400Jackson (712-165-153)
Rat IgG-HRPGoat1:400Santa Cruz (sc-2303)

Cell count

To determine the number of EGFP+ cells in P0-Cre/EGFP embryos at E11.5, two frozen sections (30 μm) at the middle level of the telencephalon were analyzed per embryo (n = 3). For the analysis of Cre expression in EGFP+ cells, eight sections per animal were examined in E9.5, E10.5, and E11.5 embryos (n = 3, respectively).

BrdU labeling

Short pulse-labeling of proliferating cells with bromodeoxyuridine (BrdU) was performed as previously described with minor modifications (Nonomura et al. 2010). Two hours following the BrdU injection (50 mg/kg) into the abdominal cavity of pregnant mice, the embryos were fixed with 4% PFA/PBS for 1 h at 4°C. To detect BrdU-incorporated cells, frozen sections (30 μm) were treated with a 2 N HCl (Wako) solution for 15 min at 37°C and neutralized in Tris-buffered saline containing 0.1% Tween 20 (Wako). To examine BrdU incorporation, the numbers of BrdU+ cells were determined in four NG2+-cell aggregates from three embryos at E11.5 using a laser scanning confocal microscopy (510 META, Carl Zeiss).

In situ hybridization

In situ hybridization of frozen sections (12 μm) was performed as previously described (Takahashi & Osumi 2002). The dissected embryos were fixed in 4% PFA/PBS overnight at 4°C. Digoxigenin (DIG)-labeled antisense and sense riboprobes for Cre recombinase were synthesized via in vitro transcription using a DIG RNA labeling mix (Roche) and T3 or T7 RNA polymerase (Promega). The probes used in this study were designed from a Cre-expressing vector (Gu et al. 1993).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

P0-Cre labeling is NCDC-specific

To obtain P0-Cre/EGFP mice for tracing the NC lineage that invade the telencephalon, we crossed the P0-Cre mouse line, in which Cre is expressed by migrating NCDCs during early embryogenesis (Yamauchi et al. 1999), with CAG-CAT-EGFP reporter line (Kawamoto et al. 2000) (Fig. 1A). We previously demonstrated the specificity of P0-Cre activity in the NC lineage in P0-Cre/EGFP mice by labeling NCDCs with the fluorescent dye DiI and revealed that these DiI-labeled NCDCs expressed EGFP in the craniofacial region (Kanakubo et al. 2006). Here, we further validated that EGFP+ cells originate from NC by checking expression of Cre mRNA and protein. In situ hybridization on sections of the telencephalon at E9.5 showed specific expression of Cre mRNA in the craniofacial mesenchyme (Fig. 1B–E). Similarly, Cre protein expression was strong in the craniofacial mesenchyme at E9.5, gradually fading out by E11.5, while these Cre-negative cells did express EGFP protein (Fig. 1F–H″). These results suggest that cells express EGFP in E9.5 and afterwards are specifically originated from the NC after Cre recombination by E9.5. Notably, EGFP+ cells located in the telencephalon did not express Cre protein (insets in Fig. 1F–H″). These findings provide strong evidence to use the P0-Cre/EGFP mice for tracing NCDCs in the telencephalon.

image

Figure 1. Strategy and specificity of the genetic labeling of neural crest-derived cells (NCDCs) using P0-Cre/EGFP mice. (A) Enhanced green fluorescent protein (EGFP) expression is induced after excision of the stop sequence (CAT/poly-A) by Cre recombination in P0-Cre/EGFP mouse (right photograph). (B–E) In situ hybridization with Cre antisense (B, C) or sense (D, E) probe on the coronal sections at the telencephalon level of the E9.5 P0-Cre/EGFP mouse. High-magnification images of the dashed squares in B and D are shown in C and E, respectively. Cre mRNA is detected in the NCDC-containing mesenchyme (B, C). (F–H′′) Expression of Cre protein in the P0-Cre/EGFP mouse telencephalon at E9.5–11.5 on the coronal sections. Each inset indicates high-magnification image of the area surrounding with the square and was taken from Z-stack images of optical sections. Cre protein is detected in EGFP+ cells in the craniofacial mesenchyme but not in the telencephalon. Cre protein in the mesenchyme gradually disappears in the EGFP+ cells between E10.5 (G–G′′) and E11.5 (H–H′′). AS, antisense probe; die, diencephalon; mb, midbrain; S, sense probe; tel, telencephalon; Tg, transgenic. Scale bars: D (for B, D), E (for C, E), 100 μm; F′′ (for F–F′′), 20 μm; G′′ (for G–G′′), H′′ (for H–H′′), 40 μm.

Download figure to PowerPoint

NCDC invasion into the telencephalon is concomitant with angiogenesis

Using P0-Cre/EGFP mice, we first analyzed when and how NCDCs invade the telencephalon through the neuroepithelial basement membrane (Fig. 2A–F). Between E9.5 and E11.5, laminin expression was observed in the basement membrane covering the telencephalic wall and blood vessels. We found that laminin+ blood vessels invaded the telencephalon in a ventral-to-dorsal gradient (arrowhead in Fig. 2A–C), confirming a previous observation (Vasudevan et al. 2008). At E9.5, a subset of EGFP+ cells were observed independent of laminin+ blood vessels (Fig. 2A–A”, D). However, the processes of EGFP+ cell were frequently observed to associate with an Flk1+ endothelial cell. Both these EGFP+ cell and endothelial cell were observed to invade the telencephalon by elongating their processes across the laminin+ neuroepithelial basement membrane (Fig. 2G–H′″). At later stages, EGFP+ cells were observed in the dorsal telencephalon, that is, at the leading front of the growth of new blood vessels (angiogenesis) (Fig. 2B–C″, arrowheads). Here, EGFP+ cells were in close contact with laminin+ blood vessels (Fig. 2E, F). These results indicate that the NCDC invasion into the telencephalon is concomitant with angiogenesis.

image

Figure 2. Penetration and distribution of neural crest-derived cells (NCDCs) in the P0-Cre/EGFP mouse telencephalon. (A–C′′) Immunostaining with anti-laminin and green fluorescent protein (GFP) antibodies on coronal sections at the telencephalic level. Laminin (green, shown in false color) is observed in the basement membrane covering the telencephalon and in the blood vessels therein. EGFP+ cells are detected by diaminobenzidine (DAB) staining to obtain high contrast images. White and black arrowheads in A–C” indicate the leading front of the growth of new blood vessels (angiogenesis). Note that a ventral-to-dorsal gradient of angiogenesis proceeds within the telencephalon between E9.5 and E11.5. Asterisks show neuroepithelial cell-like EGFP+ cells in the dorsal telencephalon. (D–F) High-magnification images of the dashed squares in A′′–C′′. A small number of EGFP+ cells are observed independent of blood vessels at E9.5 (D), although most EGFP+ cells are associated with blood vessels in later stages (E, F). (G–H′′′) A coronal section at the telencephalic level of the E9.5 embryo is stained with anti-laminin, Flk1, GFP antibodies and counter-stained 4′, 6-diamidino-2-phenylindole (DAPI). High-magnification images within the squares in G–G′′′ are shown in H–H′′′. Through the laminin+ basement membrane (H), an endothelial cell (white dashed line in H′) invades the telencephalon together with an EGFP+ cell (yellow line in H′′). Scale bars: A (for A–A′′), B (for B–B′′), C (for C–C′′), 200 μm; D–F, G′′′ (for G–G′′′), 50 μm; H′′′ (for H–H′′′), 25 μm.

Download figure to PowerPoint

NCDCs differentiate into telencephalic pericytes by E11.5

The presence of NCDCs in the mouse telencephalon has previously been reported using an Ht-PA-Cre line (Pietri et al. 2003), although it is unknown what cell types these NCDCs are. In this study, we analyzed NCDCs around/in the telencephalon by examining the expression of two markers that have been used to identify pericytes, that is, the tyrosine kinase receptor PDGFRβ (Lindahl et al. 1997) and the chondroitin sulfate proteoglycan NG2 (Ozerdem et al. 2001; Song et al. 2005). It should be noted that no markers are absolutely specific only for pericytes. PDGFRβ generally labels NCDCs in the craniofacial mesenchyme (Shinbrot et al. 1994). NG2 is expressed not only by pericytes but also by oligodendrocyte precursor cells (OPCs). However, NG2 expression in OPCs begins from E16 in the mouse forebrain (Nishiyama et al. 1996; Diers-Fenger et al. 2001). Therefore, we do not need to consider potential expression in OPCs in the stages we examined (E9.5–11.5). Morphologically, pericytes that surround endothelial cells share the basement membrane with endothelial cells (Krueger & Bechmann 2010). We confirmed this by triple-immunostaining for PDGFRβ or NG2, together with Flk1 (endothelial cell marker) and laminin (Fig. 3). Taken together, PDGFRβ and NG2 are considered as useful markers for pericytes in the stages we examined.

image

Figure 3. The usability of PDGFRβ and NG2 as pericyte markers in the telencephalon. (A–E′′) Immunostaining with anti-PDGFRβ, Flk1, and laminin antibodies on coronal section at the telencephalic level of E11.5 P0-Cre mouse embryo. (A–D) A PDGFRβ+ pericyte (yellow arrowheads) wraps around Flk1+ endothelial tube (white arrowheads), and these cell types are surrounded by laminin+ basement membrane. This pattern is also clearly observed in orthogonal view of the image shown in D (E–E′′). (F–J′′) Immunostaining with anti-NG2, Flk1, and laminin antibodies on coronal section at the telencephalon level of E11.5 P0-Cre mouse embryo. (F–I) A NG2+ pericyte (yellow arrowheads) wraps around Flk1+ endothelial tube (white arrowheads), and these cell types are surrounded by laminin+ basement membrane. This pattern is also clearly observed in orthogonal view of the image shown in I (J-J′′). Scale bars: D (for A–D), E (for E-E′′), I (for F-I), J (for J-J′′), 10 μm.

Download figure to PowerPoint

At E9.5, PDGFRβ was expressed in the craniofacial mesenchyme consisting with many EGFP+ cells, but it was not detected in the telencephalic neuroepithelium (Fig. 4A–A″). At this stage, a small number of EGFP+ cells had already invaded the telencephalon, but they were negative for PDGFRβ (insets in Fig. 4A–A″). At E10.5, PDGFRβ expression gradually became intense in regions adjacent to the telencephalon, where NCDCs later develop into the cells of the meninges (Fig. 4B–B″) (Jiang et al. 2002; Gagan et al. 2007). At E11.5, a greater number of PDGFRβ+/EGFP+ cells were observed in the ventral telencephalon (Fig. 4C–D). In contrast to PDGFRβ, NG2 was completely absent from the craniofacial mesenchyme and the telencephalon at E9.5 and E10.5 (Fig. 4E–F″). At E11.5, NG2 expression was detected in the mesenchyme around the telencephalon as well as in the telencephalic neuroepithelium (Fig. 4G–H). We confirmed this NG2 expression in the mesenchyme by double-staining with laminin (data not shown). These PDGFRβ and NG2 expression data suggest that NCDCs around/in the telencephalon differentiate into pericytes by E11.5.

image

Figure 4. Localization of pericyte marker proteins in the telencephalic regions of P0-Cre/EGFP mouse embryos. (A-D) Immunostaining with anti-green fluorescent protein (GFP) and PDGFRβ antibodies on the coronal sections of the E9.5–11.5 embryos. Insets in A-A” show high-magnification of squares in A-A′′. PDGFRβ is uniformly expressed in the craniofacial mesenchyme, including EGFP+ cells from E9.5, and this expression gradually accumulates in the region adjacent to the ventral telencephalon between E10.5 and E11.5 (B′, C′). (D) A high-magnification image within the squares in C, C′. EGFP+ cells located in the telencephalon express PDGFRβ. (E-H) Immunostaining with anti-GFP and NG2 antibodies on the coronal sections of the E9.5–11.5 embryos. NG2 is undetectable in the craniofacial mesenchyme and telencephalon until E10.5 (E′, F′). (G-G′) At E11.5, NG2 is expressed not only in the mesenchyme around the telencephalon but also in the telencephalon. (H) A high-magnification image within the squares in G, G′. Scale bars: A′′ (for A–A′′), B′′ (for B–B′′), C′ (for C, C′), E′′ (for E–E′′), F′′ (for F–F′′), G′ (for G, G′), 100 μm; D, H, 20 μm.

Download figure to PowerPoint

A small population of NCDCs retain their naive character

At E11.5, most of the EGFP+ cells expressed both PDGFRβ and NG2, and almost all NG2+ cells expressed PDGFRβ (Fig. 5A–B). Given that PDGFRβ is expressed in both pericytes and immature hematopoietic cells, such as erythroid and myeloid precursors (open arrowheads in Fig. 5A′–B; see Yoon et al. 2000), NG2 is a better marker of NC-derived telencephalic pericytes than PDGFRβ in the early embryonic stages. Therefore, we used only NG2 for the pericyte marker in the following experiments.

image

Figure 5. Existence of non-pericyte enhanced green fluorescent protein (EGFP+) cells in the telencephalon of E11.5 P0-Cre/EGFP mice. (A-B) A coronal section at the telencephalic level is stained with anti-GFP, PDGFRβ, and NG2 antibodies. Most of EGFP+ cells express both PDGFRβ and NG2 (yellow arrowheads). (B) High-magnification of the dashed square in A′′. PDGFRβ is also expressed in hematopoietic cells (open arrowheads in A′–B). (C-C′′′) A coronal section at the telencephalic level is stained with anti-GFP, NG2, and Flk1 antibodies. Although most of EGFP+ cells express NG2 (yellow arrowheads), a small population of EGFP+ cells express Flk1 (arrows) or express neither NG2 nor Flk1 (white arrowheads). (D) The ratios of NG2 and Flk1 expression within EGFP+ cells located in the telencephalon of P0-Cre/EGFP mice at E11.5. (E–E′′′) A coronal section at the telencephalic level is stained with anti-GFP, NG2, and p75 antibodies. Not only EGFP+ cells in the mesenchyme (arrows), but also NG2+/EGFP+ cells around/in the telencephalon (yellow arrowheads) express p75, an undifferentiated neural crest-derived cells (NCDCs) marker. White arrowheads show NG2+/p75+/EGFP- cells. Dotted lines indicate the border of between the craniofacial mesenchyme and telencephalic neuroepithelium. (F–F′′′) A coronal section at the telencephalic level is stained with anti-GFP, p75, and NG2 antibodies. A small number of EGFP+ cells that locate separately from blood vessels express p75 (yellow arrowheads) or express neither NG2 nor p75 (white arrowheads). (G) The ratios of p75 expression within EGFP+ cells located in the telencephalons of P0-Cre/EGFP mice at E11.5. cm, craniofacial mesenchyme; GE, ganglionic eminence; tel, telencephalon. Scale bars, A′′ (for A–A′′), B, 50 μm; C′′′ (for C–C′′′), E′′′ (for E–E′′′), F′′′ (for F–F′′′), 20 μm.

Download figure to PowerPoint

Although most of EGFP+ cells expressed NG2 (yellow arrowheads in Fig. 5C–C′′′), we sometimes observed NG2-negative EGFP+ cells that expressed Flk1 (arrows in Fig. 5C–C′′′), and those expressed neither NG2 nor Flk1 (white arrowheads in Fig. 5C–C′′′). A previous fate mapping study using P0-Cre/EGFP mice has revealed that P0 mRNA is endogenously expressed in postnatal retinal angioblasts, that is, non-NC-derived endothelial precursor cells (Kubota et al. 2011). Therefore, we quantitatively analyzed the ratio of NG2+/EGFP+ cells and Flk1+/ EGFP+ cells against EGFP+ cells at E11.5. Among the 452 EGFP+ cells observed in three embryos, 66.0% were positive for only NG2, 20.1% were positive for only Flk1, and 13.9% were negative for both NG2 and Flk1 (Fig. 5D). We carefully examined whether EGFP+ cells that expressed both NG2 and Flk1 existed or not using a confocal microscope along the z-axis, but EGFP+ cells expressing both NG2 and Flk1 were undetectable. Therefore, expression of NG2 and Flk1 is specific to pericytes and endothelial cells, respectively.

Since approximately 14% of EGFP+ cells were neither pericytes nor endothelial cells, we next examined whether or not these EGFP+ cells express p75, an undifferentiated NCDC marker (Stemple & Anderson 1992). First, we observed that the majority of NCDCs in the craniofacial mesenchyme expressed p75 at E11.5 (arrows in Fig. 5E–E′′′). Surprisingly, a large number of NG2+/EGFP+ cells that had already invaded the telencephalon also expressed p75, as well as NG2+/EGFP+ cells in the mesenchyme around the telencephalon (yellow arrowheads in Fig. 5E–E′′′). We further examined whether or not EGFP+ cells that were negative for NG2 and locating away from blood vessels express p75 in the telencephalon at E11.5. Although there existed EGFP+ cells that were negative for both NG2 and p75 (white arrowheads in Fig. 5F–F′′′), a small number of EGFP+ cells were negative for NG2 but positive for p75 (yellow arrowheads in Fig. 5F–F′′′). We quantified the latter NG2-/p75+/EGFP+ cells in the E11.5 telencephalon. Among 503 EGFP+ cells in three embryos, 94.0% were negative for p75, but 6.0% were positive for p75 (Fig. 5G). Therefore, the majority of the EGFP+ cells in the telencephalon differentiate into pericytes, though approximately 6% of EGFP+ cells seem to retain an undifferentiated NCDC character. We thus assumed that 8% of EGFP+ cells in the telencephalon that were negative for NG2 and p75 were considered to be unidentified cells.

Negative contribution of mesodermal cells in the telencephalon

Neural crest-derived pericytes have been observed in the avian forebrain within the meninges, which covers the brain, whereas mesodermal-derived pericytes have been observed within the meninges of the rest of the central nervous system in a mutually exclusive manner (Etchevers et al. 2001). That is, the forebrain (telencephalon and diencephalon) is the only part of the central nervous system to which NCDCs penetrate. Supporting this finding, a previous fate mapping study using Mesp1-Cre/R26R mice revealed that mesoderm-derived cells contribute to the mesenchyme that surrounds the midbrain (Yoshida et al. 2008). Therefore, we examined whether mesodermal-derived pericytes are present in the telencephalon using Mesp1-Cre/R26R mice. No LacZ+/NG2+ cells were observed in the telencephalon of Mesp1-Cre/R26R mice at E11.5 (n = 256 cells from two embryos, figures not shown). This finding implies that telencephalic pericytes originate from the NC rather than from the mesoderm. Alternatively, only a very small number of NCDCs were observed in the peri-midbrain mesenchymal regions at E10.5 in P0-Cre/EGFP mice (data not shown). Therefore, we assume that, in both mice and birds, mesoderm-derived pericytes occupy the midbrain.

Taken together, these data suggest that (i) a majority of NCDCs located in the telencephalon differentiate into pericytes and (ii) telencephalic pericytes are derived not from the mesoderm but from the NC.

NCDCs proliferate and differentiate into pericytes in the mesenchyme around the telencephalon

As described above, restricted NG2 expression was observed in the telencephalon and in the adjacent craniofacial mesenchyme at E11.5 (Fig. 4G′). This finding raised the possibility that NCDCs differentiate into pericytes in the telencephalon or within the mesenchyme around the telencephalon. Therefore, we examined in which regions NG2 expression begins.

At E11.0, only faint NG2 staining was detected in the mesenchyme around the telencephalon, where EGFP+ cells associate with eNOS+ blood vessels penetrating into the telencephalon (Fig. 6A–B″). Using serial sections, we observed that these EGFP+ cells also express PDGFRβ (data not shown). To confirm that NG2 expression was restricted to the mesenchyme around the telencephalon in later stages (E11.5), we immunostained thick slices (1 mm) for NG2 or the endothelial cell marker PECAM1 (Albelda et al. 1991). In the regions where large PECAM1+ blood vessels penetrate into the telencephalon, NG2+ cells accumulated at the base of the branching points of these vessels (white arrowheads in Fig. 6C, D). This was also clearly observed in triple-immunostaining for EGFP, NG2, and Flk1 (yellow arrows in Fig. 6E–E″′, Fig. 6F–F′′′). Cross-sectional views of confocal images indicated that NG2+/EGFP+ cells cover the entirety of the Flk1+ blood vessels that penetrate into the telencephalon (Fig. 6G–G″). We further investigated the proliferation of these aggregating NG2+/EGFP+ cells that surrounded the blood vessels in the mesenchyme around the telencephalon (Fig. 6H–H″). Using 2-h pulse BrdU labeling, we observed a large number of aggregating NG2+/EGFP+ cells (yellow arrowheads in Fig. 6H) that incorporated BrdU (yellow arrowheads in Fig. 6H′, H′′). Quantitatively, 31.9% of the NG2+ cells incorporated BrdU (38/119 cells, four cell aggregates from three embryos). Among these NG2+/BrdU+ pericytes, 76.3% also expressed EGFP (29/38 cells). These data indicate that NCDCs greatly contribute to the population of proliferating pericytes in the mesenchyme around the telencephalon. These results suggest a scenario that (i) differentiation of NCDCs into pericytes occurs in the mesenchyme around the telencephalon and (ii) NCDCs proliferate around the blood vessels that penetrate into the telencephalon.

image

Figure 6. Neural crest-derived cells (NCDCs) differentiate into pericytes by interacting with blood vessels. (A–B′′) Immunostaining with anti-green fluorescent protein (GFP), NG2, and eNOS antibodies on coronal section of E11.0 P0-Cre/EGFP embryo. Images in B–B′′ indicate high-magnification of the square in A–A′′. NG2 expression begins in the mesenchyme around the telencephalon and a small number of NG2+/EGFP+ cells (yellow arrowheads) are associated with eNOS+ endothelial cells (white arrowheads). (C, D) Immunostaining with anti-PECAM1 and NG2 antibodies on coronal thick slices of E11.5 P0-Cre/EGFP embryo. The large blood vessels (arrowheads in C) within the craniofacial mesenchyme branch into the telencephalon and the craniofacial mesenchyme. NG2 is strongly expressed at the base of the branching points of these large blood vessels (arrowheads in D). NG2 is also expressed in the cartilage primordium (white arrow in D) and in the blood vessels that penetrate into the eyes (black arrows in D). (E–F′′′) Immunostaining with anti-GFP, NG2, and Flk1 antibodies on floating coronal section of E11.5 P0-Cre/EGFP embryo. NG2+ cells assemble at the bases of the blood vessels that penetrate into the telencephalon (arrowheads). (F–F′′′) High-magnification of the regions indicated by arrowheads in E–E′′′ that were taken from Z-stack images of optical sections. (G-G′′) Orthogonal views of F′′′. Most of the NG2+/EGFP+ cells are wrapped around Flk1+ blood vessels. (H–H′′) The proliferation of NG2+-cell aggregates in P0-Cre/EGFP mice at E11.5. EGFP+ cells that express NG2 incorporate BrdU (yellow arrowheads in H–H′′). White arrowheads show NG2+/BrdU+/EGFP- cells. cm, craniofacial mesenchyme; tel, telencephalon. Scale bars, A′′ (for A–A′′), F′′′ (for F–F′′′), G, H′′ (for H–H′′), 50 μm; B′′ (for B–B′′), 25 μm; D (for C, D), E′′′ (for E–E′′′), 200 μm.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

In this study, we analyzed in detail the cephalic NCDC penetration into the telencephalon and differentiation into pericytes in early mouse development, which is illustrated in Fig. 7. We frequently observed that NCDCs associate with blood vessels located in the telencephalon, as described in previous reports (Etchevers et al. 2001; Korn et al. 2002; Pietri et al. 2003). Until now, it is obscure exactly when and how NCDCs invade the telencephalon. Our data demonstrated that prior to the formation of blood vessels in the neuroepithelium at E9.5, the first invasion of NCDCs into the ventral telencephalon occurred in association with invading endothelial cells. This event appears to be similar to the phenomena observed in E2 chick (Etchevers et al. 2001) and E9.5 mouse (Yoshida et al. 2008) that NCDCs associating with mesodermal-derived angioblasts form a pre-endothelial meshwork within the craniofacial mesenchyme.

image

Figure 7. A summary of the invasion and differentiation of neural crest-derived cells (NCDCs) around/in the telencephalon. At E9.5, undifferentiated NCDCs uniformly express the cephalic mesenchymal marker PDGFRβ in the craniofacial mesenchyme. The NCDCs turn off PDGFRβ expression when they begin to invade the telencephalon through the basement membrane with endothelial cells. PDGFRβ expression in the craniofacial mesenchyme gradually becomes intense in the region adjacent to the telencephalon during E10.5 and E11.5. At E10.5, a subset of NCDCs invades into the telencephalon along with blood vessels. At E11.0, NCDCs begin to differentiate into pericytes as determined by their expression of PDGFRβ and NG2. At E11.5, NG2 expression becomes strong, and NG2+ NCDCs are observed to proliferate around the angiogenic sprout in a PDGFRβ-dependent manner (also see Hellström et al. 1999).

Download figure to PowerPoint

Neural crest-derived cells penetrating into the telencephalon at E9.5 were not covered by the basement membrane (Fig. 2D, 2G–H′′′). A previous report indicates that under conditions of traumatic brain injury in the adult rat, approximately 40% of pericytes migrate away from their microvasculature and remain in a perivascular location (Dore-Duffy et al. 2000). Under the normal condition, however, it is thought to be difficult for pericytes to detach from endothelial cells because both pericytes and endothelial cells are completely covered by the basement membrane (Krueger & Bechmann 2010). Thus, it may be restricted to be only limited developmental stages (i.e. before the vascular wall is covered by the basement membrane) that a small number of NCDCs can separate from the endothelial cells.

At E9.5, NC-derived craniofacial mesenchymal cells expressed PDGFRβ as a mesenchymal cell marker, but NCDCs located in the telencephalon were negative for this marker. At E11.5, not only NCDCs located in the mesenchyme, but also NC-derived pericytes around/in the telencephalon expressed p75, a marker for undifferentiated NCDCs. Previously, it is proposed that undifferentiated perivascular mesenchymal cells that assemble around the vascular wall become vascular smooth muscle cells or pericytes, and that these differentiated cells are thought to subsequently proliferate and migrate along new angiogenic sprouts in a PDGFRβ-dependent manner (Hellström et al. 1999). Thus, it is suggested that NCDCs lose their mesenchymal character after invasion into the telencephalon, but still retain their pre-mature state even though after differentiation into pericytes (also see below).

We observed that NG2 expression started from E11.0 in NCDCs that were closely associating with endothelial tubes invading the telencephalon from the surrounding mesenchyme. Therefore, NC-derived undifferentiated perivascular mesenchymal cells may differentiate into pericytes not in the telencephalon but in the near-by mesenchyme. Although a large number of NCDCs differentiated into pericytes, approximately 14% of NCDCs located in the telencephalon did not differentiate into pericytes at E11.5, suggesting that cephalic NCDCs located in the telencephalon may be composed of heterogeneous fate-committed cells rather than of homogenous multipotent stem/precursor cells. Our findings also support the previous notion that environmental cues encountered by NCDCs may regulate their cell fates (Le Douarin & Kalcheim 1999).

In this study, we noticed unique patterns of NG2 expression in NC-derived pericytes. NCDCs around/in the telencephalon exhibited a dramatic increase in NG2 expression between E10.5 and E11.5 (Fig. 4F′, G′). Our BrdU data indicate that approximately 30% of NG2+ cells in the mesenchyme around the telencephalon incorporate BrdU, 2 h after a BrdU injection at E11.5. In association with this finding, a previous study reported that nearly half of the PDGFRβ+ pericytes in the brain are labeled with BrdU under the same conditions (Hellström et al. 1999). Thus, pericyte proliferation is maintained and may be increased after invasion into the telencephalon at E11.5. This also explains why NG2 expression dramatically increases around/in the telencephalon and why NC-derived pericytes maintain p75 expression.

It is of note that there was a time lag between the invasion of NCDCs into the telencephalon from E9.5 and their differentiation into pericytes around E11.5. Similar observation has previously been reported in a quail-chick chimeras study showing that NCDCs, which have invaded the forebrain from E4, express a pericyte marker 4 days later, that is, at E8 (Etchevers et al. 2001). We thus suspect that these pioneering NCDCs located in the telencephalon might remain undifferentiated.

From our observation, approximately 14% of EGFP+ cells in the E11.5 telencephalon did not express pericyte markers. What are these non-pericyte EGFP+ cells? There are several possibilities. First, they are not derived from the NC (i.e. negative for p75) but may ectopically express Cre. However, we believe that this possibility is unlikely, given that the expression of Cre protein was nearly absent by E11.5 in the telencephalic region of P0-Cre/EGFP mice (Fig. 1F–H′′). Second, considering that many of NCDCs retain p75 expression after differentiation into pericytes, a part of NCDCs that do not differentiate into pericytes may lose their p75 expression after penetration into the telencephalon. In the chick, a previous study has shown that NC-derived neurogenic precursors that occupy a minor population in the lateral migration pathway are eliminated, probably by localized environmental cues (Wakamatsu et al. 1998). Therefore, one possibility is that p75-/EGFP+ cells (approximately 8% of NCDCs) in the E11.5 telencephalon might eventually be removed from the telencephalon. Third, a recent study using Wnt1-Cre mouse line demonstrated that approximately one-third of gonadotropin-releasing hormone-1-positive (GnRH-1+) neurons in the postnatal forebrain are derived from the NC (Forni et al. 2011). However, these GnRH-1+ forebrain neurons are detectable from E12.5 (Wray et al. 1989). It is therefore improbable that non-pericyte EGFP+ cells located in the telencephalon at E11.5 are GnRH-1+ neurons. Fourth, NCDCs that re-invade the telencephalon may become neuroepithelial cells under environmental influences. This idea is based on the fact that we occasionally observed a small number of EGFP+ cells that morphologically resembled neuroepithelial cells in the dorsal telencephalon at E9.5 (data not shown). These neuroepithelial-like EGFP+ cells were also observed at E10.5 and E11.5 (see asterisks in Fig. 2B′, B′′ and Fig. 2C′, C′′). There are no good markers to determine the molecular character of these EGFP+ cells as neuroepithelial cells, but the incidence of these neuroepithelial-like EGFP+ cells was very low (3/10 embryos). That is, after the epithelial-mesenchymal transition, only a small number of NCDCs may enter the neural tube to become neuroepithelial cells. This scenario would therefore occur only rarely. Fifth, p75+/EGFP+ cells locating away from blood vessels differentiate into pericytes after penetrating into the telencephalon. As described above, however, pericyte differentiation appears to occur in the mesenchyme around the telencephalon. Currently, although it remains controversial, we favor an interpretation that p75+/EGFP+ cells may retain undifferentiated NCDC characteristics within the telencephalon. Future studies using Cre-inducible fate mapping systems, such as NC lineage tracing with the Sox10-codon-improved CreERT2 mouse line (Shimshek et al. 2002; Simon et al. 2012) would provide a better understanding features of NCDCs in the brain.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

We thank Dr Kenichi Yamamura for the P0-Cre mice, Dr Junichi Miyazaki for the CAG-CAT-EGFP mice and Dr Anastassia Stoykova for a Cre-containing plasmid. We are grateful to Ms Ayumi Ogasawara-Shirotori, Ms Sayaka Makino and Ms Emi Ootsuki for animal care and technical assistance. We thank Dr Yoshio Wakamatsu and all of the other members of the Osumi laboratory for their valuable comments and discussion. This work was supported by the Global COE Program “Basic and Translational Research Center for Global Brain Science” of MEXT (to N.O.). E.Y. was a research associate of the Global COE program.

Author contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References

E.Y. designed research, performed all experiments, analyzed data, and wrote the manuscript. M.T. designed research, analyzed data, and wrote the manuscript. Y.S. provided Mesp1-Cre/R26R mice. N.O. designed research, analyzed data and wrote the manuscript. All authors read and approved the final manuscript.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. Author contributions
  9. References