Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Feb 2.
Published in final edited form as: Dev Dyn. 2006 Dec;235(12):3336–3347. doi: 10.1002/dvdy.20988

The Initial Fetal Human Retinal Vasculature Develops by Vasculogenesis

D Scott McLeod 1, Takuya Hasegawa 1, Tarl Prow 1, Carol Merges 1, Gerard Lutty 1,*
PMCID: PMC3271053  NIHMSID: NIHMS215617  PMID: 17061263

Abstract

There is increasing evidence that the hemangioblast, a common progenitor for hematopoietic cells and endothelial cells, participates in embryonic and extra-embryonic vasculogenesis in some organs. Whether resident angioblasts or endothelial progenitor cells (EPCs) contribute to human retinal vasculogenesis is still a matter of controversy. To address this controversy, fetal human retinas of 6–23 weeks gestation (WG) were examined using immunohistochemistry and a panel of antibodies against endothelial cell markers (CD34, CD31), a marker for retinal angioblasts and endothelium (CD39/ecto-ADPase), and a marker for precursors and hemangioblasts (CXCR4). Confocal microscopic spectral analysis and double labeling with Ki67 was used to identify the proliferating cell types. In the inner neuroblastic layer of the 6–8 WG retina and in the putative ganglion cell layer in avascular regions of older eyes (14 WG–20 WG), scattered CD39+ angioblasts were well in advance of forming vasculature. There was a layer of CXCR4+ cells in the inner retina that was reduced in size with development. As blood vessels formed, CD39+ cells were always well in advance of the vascular front and they expressed CXCR4. This demonstrates that a pool of resident angioblasts express CD39 and CXCR4 as they differentiate and participate in vasculogenesis in the fetal human. They retain expression of CD39 as endothelial cells in the newly formed retinal vasculature but they down-regulate CXCR4 expression.

Keywords: progenitors, vasculogenesis, angioblasts, endothelial cells, astrocyte precursor cells, astrocytes, proliferation

Introduction

Two modes of blood vessel assemblage, vasculogenesis and angiogenesis, may contribute to retinal vascularization in humans. Vasculogenesis is the de novo formation of blood vessels from mesenchymal precursors or angioblasts, whereas angiogenesis is the formation of blood vessels by proliferation and budding from pre-existing vessels. Vasculogenesis has been described in other developing organ systems like heart, lung, spleen, and pancreas (Risau and Flamme, 1995; Beck and D'Amore, 1997; Risau, 1997, 1998) and to a much a lesser degree in retina (McLeod et al., 1987; Hughes et al., 2000; Chan-Ling et al., 2004). The apparent insufficient evidence for retinal vasculogenesis is due to a lack of a specific marker or markers for angioblasts. It was Ashton who originally suggested that, in fetal human, mesenchymal precursors existed in a vanguard that preceded the ingrowth of retinal blood vessels (Ashton, 1970). We have described them previously in dog (McLeod et al., 1987; Lutty and McLeod, 1992; McLeod et al., 1996a) and more recently in human (Chan-Ling et al., 2004).

The time course of human retinal vascular development has been known for nearly a century (Mann, 1928) and has continued to be studied as new techniques evolve. A number of elegant immunohistochemical whole-mount studies have examined different cell types in relationship to blood vessel development and suggested that there is an intimate concomitant relationship between the astrocytes and forming blood vessels (Chan-Ling et al., 2004). Sandercoe et al. used double labeling with CD34/Ki67 and GFAP/Ki67 in whole-mounts and found that, at 18 weeks gestation, the majority of proliferating cells associated with developing vessels were astrocytes (Sandercoe et al., 1999). More recently, astrocyte precursor cells (APCs) have been described in developing fetal human retina as well as a population of angioblasts (Chu et al., 2001; Chan-Ling et al., 2004). All of theses studies have added considerably to our knowledge of retinal vasculogenesis. Unfortunately, the bulk of this work has been done on whole-mounts where only one or two markers can be employed. In addition, and perhaps more importantly, the retinal angioblast has proven to be somewhat elusive to identify and study in that angioblast-specific markers do not exist. In order to identify angioblasts and distinguish them from endothelial cells, a battery of markers must be used in serial sections.

In past studies, we have used ADPase enzyme histochemistry, which is now known to be CD39, as a marker for precursors (angioblasts) before, during, and after their differentiation and organization into primordial blood vessels. It is important to note that the staining pattern seen in whole-mount preparations was identical using both ADPase enzyme histochemistry (Lutty and McLeod, 1992; Chan-Ling et al., 2004) and CD39 immunohistochemistry in the fetal human as a marker for angioblasts and blood vessels. It is present in adult retinal endothelial cells of all species studied to date and we have used it to study pathological neovascularization in sickle cell and diabetic retinopathy as well as retinopathy of prematurity (Lutty and McLeod, 1992, 2005; McLeod et al., 1993, 1996b, 1997; Lutty et al., 1994, 1996). CD39, or nucleoside triphosphate diphosphohydrolase-1 (NT-PDase1), is the major vascular endothelial ecto-nucleotidase and hydrolyses nucleoside triphosphates and diphosphates, ultimately to the nucleoside analogues (Goepfert et al., 2001). Hypoxia induces expression of CD39 in human microvascular cells (Eltzschig et al., 2003) and CD39′s ability to inhibit platelet aggregation is lost in oxidative stress (Kaczmarek et al., 1996). Therefore, CD39 (ADPase) has been the only marker to date for labeling angioblasts from our studies of vasculogenesis in dog and man.

CXCR4 is a 352 amino acid protein, which displays 33% homology to CXC and CC members of the chemokine receptor family. CXCR4 is highly conserved across species and is expressed on a wide variety of cell types including hematopoietic cells, vascular endothelial cells, neurons, microglia, and astrocytes (Lazarini et al., 2003). CXCR4 is the sole receptor for the alpha chemokine stromal-derived factor 1 (SDF-1). SDF-1 engagement of CXCR4 leads to diverse biological responses, some of which are cell-type dependent and include upregulation of integrin-mediated adhesion, chemotaxis and migration, enhanced survival and proliferation, as well as apoptosis (Juarez et al., 2004). CXCR4 is upregulated in hypoxia and is expressed in canine retinal angioblasts in vitro (Lutty et al., 2006).

CD34 is a105- to 120-kD transmembrane cell surface glycoprotein, which is thought to be selectively expressed within the human and murine hematopoietic systems on stem and progenitor cells (Civin et al., 1984; Tavian et al., 1996; Labastie et al., 1998; Peichev et al., 2000). CD34 is also commonly used as a marker for endothelial cells (Fina et al., 1990). In a previous study, we found that CD34 is only expressed on endothelial cells of formed retinal vessels and not by retinal angioblasts (Chan-Ling et al., 2004).

CD31 or PECAM-1 (platelet endothelial cell adhesion molecule) is a 100-kDA glycoprotein in endothelial cells and a 130-kD glycoprotein in platelets. CD31 is an adhesion molecule that controls the diapedesis of leukocytes and transmigration of CD34+ cells across endothelial cells (Imhof and Dunon, 1995; Yong et al., 1998). CD31s homophilic binding contributes to maintenance of endothelial cell integrity (Newman, 1997). CD31 is frequently used as a marker for endothelial cells in adult tissues.

The goal of the present work was to define the temporal and spatial relationships of a variety of cell types and retinal blood vessels in a comprehensive sequential fashion in serial sections of fetal human eyes ranging from 6–23 weeks gestation. We used a battery of antibodies to label endothelial cells (CD31, CD34), angioblasts (CD39), precursors (CXCR4), smooth muscle cells and pericytes (alpha smooth muscle actin, NG-2), microglia (MHC class II), astrocyte precursor cells (Pax-2), and astrocytes (GFAP).

Additionally, we used a marker for proliferating cells, Ki67, and immunohistochemical double labeling and confocal microscopy to determine where and what type of cells were proliferating during blood vessel formation. Using this strategy, it was demonstrated that the initial superficial vasculature of human retina develops by vasculogenesis without endothelial cell or angioblast proliferation.

Results

The 6-WG retina (Fig. 1) contained no blood vessels but an extensive intravitreal vascular network was present as well as a developing choriocapillaris. Expression of endothelial cell markers (CD34/CD31/CD39) indicated that choriocapillaris was not fully contiguous (patchy labeling) but this vasculature served as a positive control for the vascular marker antibodies (Fig. 1B,C). Because the staining pattern of CD34 and CD31 was identical, only CD31-stained sections are shown in all figures. Some single scattered CD39-positive cells were located within the inner portion of the peripapillary retina (Fig. 1C), which at this stage of development consisted almost entirely of an undifferentiated neuroblastic layer with no laminar structure (Fig. 1A). There was a layer of CXCR4+ cells that occupied nearly half of the inner neuroblastic layer adjacent to the transient fiber layer of Cheivitz in the posterior pole of retina and tapered toward midperipheral retina to a single line of cells (Fig. 1E). The transient fiber layer of Cheivitz separates the outer and inner neuroblastic layers in embryonic retina and becomes the inner plexiform layer in fetal human retina (Rhodes, 1979). Neither CD45- nor MHC-class 2–positive cells were found in retina but a few were observed in vitreous and developing choroid (Fig. 1D). There were no cells expressing SMA or NG2 (smooth muscle cells or pericytes) in retina or choroid. Pax-2 and GFAP immunoreactivity were also completely negative (not shown). Proliferation, as determined by Ki67 staining, was confined to the outer portion of the undifferentiated neuroblastic layer (Fig. 1F).

Fig. 1.

Fig. 1

Open in a new tab

Sections from the developing optic nerve head region of a 6-WG embryonic eye (optic nerve to the right). A: The hematoxylin and eosin (H&E) -stained section shows a few cells in the innermost retina and within the developing nerve (arrowheads). B: CD31 was only present in the developing choriocapillaris (bottom right) (arrow indicates the bleached RPE in all panels). C: CD39 staining was observed in a few round-shaped cells in the inner neuroblastic layer and in cells in the developing nerve fiber layer (arrowheads). CD39 was also present in the developing choriocapillaris (bottom). D: CD45+ cells were scattered throughout choroid (asterisks). E: CXCR4 was observed throughout the inner neuroblastic layer and within the inner retina and developing optic nerve head (arrowheads). F: Ki67 was only associated with cells within the outermost neuroblastic layer. B-F: APase reaction product.

The 7-WG retina (Fig. 2) also was avascular and, therefore, endothelial cell markers (CD34/CD31) were negative. Some single scattered CD39-positive cells, some of which were spindle-shaped, were located within the nerve fiber layer of the peripapillary retina (Fig. 2C). Like the 6-WG retina, there were layers of CXCR4+ cells that occupied nearly half of the inner neuroblastic layer in the posterior pole and tapered toward midperipheral retina to a single layer. Near the optic nerve head, some CXCR4-positive cells were located within the nerve fiber layer, and some were spindle-shaped like the cells labeled with CD39 (Fig. 2E). A few CD45 labeled cells were also located in that region (Fig. 2D). No MHC-class 2–positive cells were found in retina. SMA, NG2, Pax-2, and GFAP staining were also completely negative (data not shown). Proliferation (Ki67+) was still confined to the outer portion of the undifferentiated neuroblastic layer (Fig. 2F). The results at 9 WG were similar for all markers and again the retina was non-vascularized (results not shown).

Fig. 2.

Fig. 2

Open in a new tab

Sections from the optic nerve head region of a 7-WG embryonic eye (optic nerve to the right). A: The H&E-stained section shows round and spindle-shaped cells in the innermost retina and within the developing nerve fiber layer (arrowheads). B: CD31 was only present in the developing choriocapillaris (arrow indicates the bleached RPE in all panels). C: CD39 was present in cells in the nerve fiber layer (arrowheads). Staining was also seen in the developing choriocapillaris. D: CD45+ cells were scattered throughout choroid and some in the vitreous cavity. E: CXCR4 was observed throughout the inner neuroblastic layer and within cells of the nerve fiber layer (arrowheads), similar to those in the H&E (A) and the cells labeled with CD39 (C). F: Ki67 was only associated with cells within the outermost portion of the neuroblastic layer of retina and within the choroid. B–F: APase reaction product.

The 12-WG retinas were negative for endothelial cell markers (CD34/CD31) and were, therefore, considered non-vascularized (Fig. 3). This differs from our prior study with the Chan-Ling lab but the gestational ages were calculated by different techniques in the two studies (Chan-Ling et al., 2004). SMA, Pax-2, and GFAP immunostaining were also completely negative (not shown). However, Pax-2+ cells were observed in the optic nerve proper (data not shown), as demonstrated previously by Chu et al. (2001). Single scattered CD39+ cells were located within the nerve fiber layer throughout the retina and at this age extended to the far periphery. Near nerve head they were spindle-shaped (Fig. 3C) and were more numerous than in the 9 WG. Again, there were layers of CXCR4+ cells that occupied nearly half of the inner neuroblastic layer in the posterior pole and tapered toward the far peripheral retina into a single layer (Fig. 3E). Near nerve head, some were located within the nerve fiber layer and had a spindle shape (Fig. 3E). A few CD45+ cells were located in the nerve fiber layer near nerve head (Fig. 3D). Some MHC-class 2+ cells were found in peripheral retina, and were more numerous than at 9 WG (data not shown). Proliferation was confined to the outer portion of the undifferentiated neuroblastic layer (Fig. 3F).

Fig. 3.

Fig. 3

Open in a new tab

Sections from the peripapillary region of a 12-WG fetal eye (optic nerve to the right and out of the field). A: The H&E-stained section shows scattered cells within the developing nerve fiber layer (arrowheads) and a demarcation between the inner and outer neuroblastic layer. B: CD31 was only present in the developing choriocapillaris (arrow indicates the bleached RPE in all panels). C: CD39 staining was observed in scattered cells in the nerve fiber layer (arrowheads) and some in the inner portion of the neuroblastic layer. Staining was also seen in the developing choriocapillaris. D: CD45 labels cells in the nerve fiber layer but the number of cells was less then those stained with CD39. E: CXCR4 was observed within the inner neuroblastic layer and in scattered cells within the nerve fiber layer (arrowhead), similar to those in the H&E (A) and the cells labeled with CD39 (C). F: Ki67 staining was only associated with cells within the outermost portion of the neuroblastic layer of retina and within the choroid. B–F: APase reaction product.

In the 14-WG retina (Fig. 4), the putative ganglion cell layer had formed in addition to the inner plexiform layer, providing a definitive separation of the outer retina that still consisted of a single undifferentiated neuroblastic layer (Fig. 4A). Retinal blood vessels extended in two arcades (superior and inferior) just beyond the confines of the optic nerve head. CD34 and CD31 were restricted to the formed retinal blood vessels only (Fig. 4B). In contrast, CD39 labeled not only formed retinal vessels, but also labeled individual cells in advance of vessels extending well into the avascular periphery (Fig. 4C). No CD34+ cells in these vessels were proliferating based on the lack of Ki67 in double-labeled sections (data not shown). CXCR4 was not evident in blood vessels in optic nerve head and the peripapillary nerve fiber layer. However, the sections immunolabeled with CXCR4 may not have contained the few retinal vessels present in that region of the eye (Fig. 4E). CXCR4 immunoreactivity was found in all cells within the putative ganglion cell layer anterior to the peripapillary region (inset in Fig. 4E). CD45+ cells continued to be rare, except near established vessels (Fig. 4D). Pax-2- and GFAP-positive cells were absent in avascular regions but were present in the vascularized peripapillary nerve fiber layer (Fig. 4G and H). Ki67-positive cells were seen in the peripapillary nerve fiber layer and in the outer portion of the neuroblastic layer (Fig. 4F). Double labeling with fluorescent probes (KI67/Pax-2) and confocal microscopy revealed that the proliferating cells were APCs (data not shown). MHC-class 2 staining was present in a few cells near nerve head but none were observed in peripheral retina, unlike the 12-WG specimen (data not shown). SMA and NG2 labeled a large vessel in nerve head (presumably an artery) and perhaps its branches in the peripapillary nerve fiber layer (data not shown).

Fig. 4.

Fig. 4

Open in a new tab

Sections from the peripapillary region of a 14-WG fetal eye at the onset of retinal vascularization (optic nerve to the right and out of the field). A: The H&E-stained section shows a well-developed nerve fiber layer, and a clearly defined inner and outer neuroblastic layer separated by the inner plexiform layer. B: CD31 was present in the developing peripapillary retinal vessels (arrow indicates the edge of formed vessels in all panels). C: CD39 staining was observed in formed retinal vessels and in scattered cells in advance. D: CD45 labeled single cells in the nerve fiber layer but the number of cells was less then those stained with CD39. E: CXCR4 was observed within the apparent ganglion cell layer in this region and not within the formed vessels. In midperipheral regions, however, CXCR4+ cells were observed throughout the inner retina (arrowhead in inset). F: Ki67 was now associated with cells in the vicinity of the developing retinal vessels and in the outer neuroblastic layer. G: Pax-2 labeling was also evident in the region of developing vessels as was GFAP staining (H). B–H: APase reaction product.

At 16 WG (Fig. 5), the retinal structure was like that of the 14-WG eye, (trilayered: outer undifferentiated neuroblastic; inner plexiform; putative ganglion cell) except the retinal vessels had begun to form a 4-lobe butterfly shape, with the superior and inferior arcades extending further into the temporal and nasal retina (Chan-Ling et al., 2004). As was the case at all ages studied, anti-CD31 and anti-CD34 immunolabeled only formed retinal vessels (Fig. 5B), while CD39 labeled both vessels and inner retinal cells well in advance of the edge of the vasculature (Fig. 5C). CXCR4 expression was not associated with formed retinal blood vessels, but some spindle-shaped cells at the edge of the vasculature were weakly labeled. All cells in the putative ganglion cell layer of avascular retina were intensely labeled. A few CD45+ cells were located in vascularized regions of retina (Fig. 5D) and occasionally a few were scattered in avascular regions. MHC class-2+ cells were rare anywhere in retina (data not shown). Pax-2+ cells were present in vascularized regions of retina but ended abruptly just anterior to the peripheral edge of the vasculature. GFAP+ astrocytes were only associated with formed retinal blood vessels, and never observed in advance of them. Ki67 staining demonstrated a decrease in proliferation in the outer neuroblastic layer posterior to the vasculature but numerous labeled cells were located in the outer neuroblastic layer of the avascular retina. Ki67+ cells were scattered within the vascularized inner retina, which like Pax-2 ended abruptly just anterior to the peripheral edge of formed vessels. Most Ki67+ cells in vascularized regions were perivascular in location. NG-2 labeled cells in vascularized retina (not shown), had a perivascular appearance (presumed pericytes). SMA staining (not shown), was restricted to some large retinal vessels (presumed arteries).

Fig. 5.

Fig. 5

Open in a new tab

Sections from the peripheral edge of formed vasculature of a 16-WG fetal eye. A: The H&E-stained section shows a well-developed nerve fiber layer, and a clearly defined inner and outer neuroblastic layer separated by the inner plexiform layer. B: CD31 was present in the formed retinal vessels (arrow indicates the edge of formed vessels in all panels). C: CD39 staining was observed in formed retinal vessels (arrow) and in scattered cells in advance (arrowhead). D: Single CD45+ cells were scattered cells in the nerve fiber layer. E: CXCR4 was observed within the apparent ganglion cell layer and to a lesser degree within spindle-shaped cells in advance of formed vessels (paired arrows). F: Ki67 was associated with cells in the vicinity of the edge of developing retinal vessels, some slightly in advance (asterisk) and in the outer neuroblastic layer. G: Pax-2 labeling was also evident at the edge of developing vessels and in cells slightly in advance (asterisk). H: GFAP staining clearly lagged behind the edge of formed vessels. B–H: APase reaction product; small arrows indicate the RPE in all panels.

In the 23-WG eye (Fig. 6), the retina still had a tri-layered structure in avascular periphery and noticeably less cells in the putative ganglion cell layer. This could simply be the result of the thinning of this and other layers because the eye has grown considerably in size. In the vascularized region, an outer plexiform layer had begun to form resulting in separation of the inner and outer nuclear layers (Fig. 6A). This was especially true in the temporal area. Rudimentary photoreceptor inner segments were also present in these regions. At the edge of formed vasculature, CD31 and CD34 only labeled endothelial cells of blood vessels (Fig. 6B). CD39+ cells were located in and around blood vessels and within single cells in the putative ganglion cell layer throughout the entire extent of avascular periphery (Fig. 6C). CXCR4 expression appeared to be associated with ganglion cells of the posterior pole, especially in the temporal aspect (presumably the forming macula). At the edge of the forming vasculature, CXCR4 was observed in spindle-shaped cells and in cells throughout the avascular inner retina (Fig. 6E). CD45+ cells were most abundant near nerve head, but there were none associated with blood vessels (Fig. 6D). Some scattered single CD45+ cells were observed in the avascular inner retina. The number of CD45+ cells, however, was far less than those expressing CD39. MHC class-2+ cells were numerous near nerve head but the number of cells steadily declined peripherally and did not extend to the edge of formed blood vessels (not shown). They were located both in the inner retina and within the definitive inner nuclear layer where it was present. Both Pax-2 and GFAP positive cells were absent in avascular regions but were found at the peripheral edge of forming vasculature and throughout the region with formed vessels posteriorly (Fig. 6G and H). Proliferating cells were abundant in the outer portion of the neuroblastic layer in avascular retina but the number diminished dramatically closer to the edge of established vasculature and was virtually absent in the posterior neural retina. In inner retina, Ki67+ cells were observed at the border of vascularized retina (Fig. 6F), and to a lesser degree more posteriorly. Again, double labeling demonstrated that the majority of these cells were APCs (Fig. 7B). NG-2+ cells surrounded most retinal vessels of all sizes, including those near the far peripheral edge of the formed vasculature (data not shown). SMA labeling was only observed in scattered large and medium size retinal vessels (presumed arteries and arterioles) and was most prominent in more established posterior vasculature (data not shown).

Fig. 6.

Fig. 6

Open in a new tab

Sections from the peripheral edge of formed vasculature of a 20-WG fetal eye. A: The H&E-stained section shows a well-developed nerve fiber and ganglion cell layers, and a clearly defined developing inner and outer nuclear layers separated by the outer plexiform layer. B: CD31 was present in the formed retinal vessels (arrow indicates the edge of formed vessels in all panels). C: CD39 was observed in formed retinal vessels and in scattered cells in advance (arrowheads). D: CD45+ cells were scattered cells in the nerve fiber layer but they are rare in this region. E: CXCR4 was observed within the apparent ganglion cell layer and to a lesser degree within spindle-shaped cells in advance of formed vessels (arrowheads). F: Ki67 was associated with cells in the vicinity of the edge of developing retinal vessels and in the outer neuroblastic layer. G: Pax-2 labeling was also evident at the edge of developing vessels and in cells slightly in advance. H: GFAP staining clearly lagged behind the edge of formed vessels. B–H: APase reaction product; small arrows indicate the RPE in all panels.

Fig. 7.

Fig. 7

Open in a new tab

Multispectral confocal analysis of Ki67 (green arrows), CD34 (red arrows in A), Pax-2 (red arrows in B), and GFAP (red arrows in C) colocalization was performed in sections from a 20-WG human fetal retina. All sections were counterstained with mounting media containing DAPI (blue). Spectral analysis revealed the presence of dual Pax-2-positive and GFAP-positive cells (yellow arrows in B,C) at the edge of forming vasculature. The majority of Pax-2-labeled cells were proliferating as determined by cell-specific spectral analysis. No CD34/Ki67 double-labeled cells were observed in these sections. D: Double labeling with Ki67 (green) and CD39 (red) in the nerve fiber layer of peripapillary region of a 14-WG retina demonstrates that adjacent cells and not angioblasts express Ki67. E: Ki67 (green) and CXCR4 (red) in the peripapillary retina of a 7-WG eye demonstrates that proliferation in this region is in the posterior region of the neuroblastic layer (double arrow), while CXCR4 is the inner portion of the neuroblastic layer and inner retina.

Double Labeling for Proliferation

Multispectral confocal analysis of double-labeled sections was performed to determine which cell types associated with vasculogenesis of the retina were proliferating. The following combinations of antibodies were used: Ki67/CD34, Ki67/Pax2, Ki67/GFAP, Ki67/CXCR4, and Ki67/CD39. Colocalization was assessed in sections from a 14-, 16-, and 20-WG human fetal retina (Fig. 7A–E). All sections were immunostained with mouse anti-Ki67 except Figure 7A and D, which were stained with rabbit anti-Ki67. All Ki67 was detected with either goat anti-mouse or goat anti-rabbit secondary antibodies conjugated to FITC (green arrows in 7A–C and E). Other antigens were detected with mouse anti-CD34 (Fig. 7A), rabbit anti-Pax2 (Fig. 7B), and rabbit anti-GFAP primary antibodies (Fig. 7C) followed by goat anti-mouse and goat anti-rabbit secondary antibodies all conjugated to Cy3 (red arrows). All sections were counterstained with mounting media containing DAPI. Sections were examined, imaged, and analysed using a Zeiss 510 META confocal microscope. Spectral analysis revealed the presence of dual-labeled Ki67/Pax-2-and Ki67/GFAP-positive cells (had spectral peaks at both fluorochrome emission wavelengths) at the edge of forming vasculature (yellow arrows in Fig. 7B,C). The majority of Pax-2+ cells were proliferating as determined by Ki67/Pax-2 dual labeling and cell-specific spectral analysis. No CD34/Ki67 double-labeled cells were observed in sections (Fig. 7A). No CD39+ cells were positive for Ki67 (Fig. 7D). Sections were also double labeled for CXCR4 and Ki67 (Fig. 7E). At ages when the retina was avascular, the majority of the cells expressing Ki67 were in the posterior portion of the neuroblastic layer, posterior to the CXCR4+ cells.

Double Labeling for CXCR4/CD39 and CXCR4/Ki67

Double labeling of cryosections and whole-mounts with anti-CXCR4 and anti-CD39 was used to determine if the pool of CXCR4+ progenitors seen vitread to the layer of Cheivitz at the earliest ages in embryonic development (6–7.5 WG), and in the putative ganglion cell layer as development progressed, gave rise to CD39+ angioblasts. Eyes from 7, 9, 12, 14, 16, 20, and 23 WG were examined, imaged, and analysed using the Zeiss 510 META confocal microscope. In the older vascularized specimens 16–23 WG, we were able to examine the avascular retina, edge of the vasculature, and vascularized regions of retina.

In the 7-WG retina, the layer of Cheivitz was prominent and all cells in the adjacent inner neuroblastic layer labeled with CXCR4. At the optic nerve head, single scattered cells expressing CXCR4/CD39 were observed (Fig. 8A–C). In older specimens, double-labeled cells in the ganglion cell layer had a rounded morphology, while others within the nerve fiber layer had a spindle shape (Fig. 8D–G). At every age examined, prior to the appearance of retinal vessels at 14 WG, we found CXCR4+/CD39+ cells distributed throughout the inner retina and many were spindle-shaped in regions where the major vascular arcades would eventually form. In older eyes, these cells were observed further towards the periphery. In the avascular periphery and at the border of vascularized retina in sections from the 16–23-WG specimens, CXCR4+ cells, with a rounded morphology, occupied the entire putative ganglion cell layer.

Fig. 8.

Fig. 8

Open in a new tab

Double labeling of embryonic/fetal retina in sections and wholemounts. A–C: Section from nerve head region of a 7-WG embryonic retina double labeled with CD39 (red) and CXCR4 (green). In A, both channels are shown while CXCR4 (B) and CD39 (C) are shown independently. The arrows in all indicate double-labeled CD39+/CXCR4+ cells, some having a spindle-shaped morphology. D–G: Section from the peripheral edge of forming vasculature in a 23-WG retina double labeled with CD39 (red) and CXCR4 (green) and counterstained with DAPI (blue). Double-labeled CD39+/CXCR4+ angioblasts (arrows in all) are shown migrating from the ganglion cell layer to the inner retina. In D, all channels are shown. In E, both the red and green channels are shown, while in F and G the red (CD39) and green (CXCR4) channels are shown independently. H,I: Retinal whole mount from a 22-WG specimen double labeled with CD39 (red) and CXCR4 (green). At the edge of forming vasculature (H), double-labeled CD39+/CXCR4+ elongated angioblasts with processes (arrows) are seen aligning with formed lumen. In the avascular retina (I), double-labeled CD39+/CXCR4+ angioblasts with a round morphology are shown (arrows)

When retinas were viewed as flat mounts, many of the CXCR4+ cells were double labeled with CD39 (Fig. 8H,I). At the border of vascularized retina, dual-labeled cells appeared migratory (Fig. 8H), moving inward into the nerve fiber layer and adjoining the tips of forming vessels, while those in avascular retina had a rounded morphology (Fig. 8I). Once CD39+ cells were incorporated into the vasculature, CXCR4 expression was down-regulated. CXCR4+ cells double-labeled with Ki67 were never observed in our specimens. At lower magnification, the CD39-positive vasculature and angioblasts coalescing to form cords were apparent (Fig. 9A). This was apparent especially in whole mounts at the advancing edge of the vasculature as well as within areas of formed vasculature, where new cords were forming to increase the density of the capillary plexus. At the tips of the forming blood vessels, cells in vascular cords were double labeled for CXCR4 and CD39 (Fig. 9B) but in the mature areas of vasculature, CXCR4 expression was down-regulated in endothelial cells.

Fig. 9.

Fig. 9

Open in a new tab

Retinal whole mount from a 22-WG specimen double labeled with CD39 (red) and CXCR4 (green). A: With only the red channel shown (CD39), the newly formed retinal vasculature is apparent. Angioblasts are assembling at the tips of the new blood vessels while many individual angioblasts are apparent in avascular retina. B: At the edge of forming vasculature, double-labeled CD39+/CXCR4+ elongated angioblasts with processes are seen aligning to form a vascular cord (arrows).

Discussion

This is the first time that a series of vascular and nonvascular immunohistochemical markers has been employed to examine vascular development from 6 WG through 23 WG in embryonic/fetal human retina. Changes in these markers were studied in the avascular retina during embryonic development and in the fetal period when maturation of blood vessels (NG-2+ and SMA+ accessory cells investing blood vessels) was evident posteriorly. We have shown in previous studies that resident angioblasts, present well in advance of developing blood vessels of inner retina, participate in the initial formation of the fetal human (Chan-Ling et al., 2004) and newborn dog retinal vessels (McLeod et al., 1987, 1996a). The fact that we never observed proliferation associated with primordial capillary formation (CD39+/Ki67+) strengthens the role for angioblastic differentiation and organization contributing to primary or superficial vascular development by vasculogenesis.

In early embryonic development in the current study, angioblasts (CD39+/CXCR4+) are present in inner retina well before a vasculature forms and these precursors appear to be derived from the pool of postmitotic progenitors (CXCR4+) in the inner neuroblastic layer. Many of the angioblasts are spindle shaped and appear to be migrating from the pool of CXCR4+ in the inner neuroblastic layer (Fig. 8A–G). In our prior studies using ectoADPase labeling alone in neonatal dog and fetal human, this is exactly the disposition of the angioblasts as they appear to migrate posteriorly to aggregate and assemble into vascular cords (Chan-Ling et al., 2004; Lutty and McLeod, 2003). In this study, we did not find CXCR4+ cells proliferating (double labeled with Ki67) at any time point examined. As development progresses and the laminar structure of the retina becomes apparent, CXCR4+ cells are restricted to the inner retina where primordial blood vessels will form. Many of these cells in avascular peripheral retina co-express CXCR4 and CD39 before and during assembly at the tips of forming vasculature. As apparent lumenization occurs, CXCR4 expression is down-regulated, and is not associated with more mature posterior retinal blood vessels. It is quite possible that these CXCR4+ progenitors are multipotent and probably give rise to nonvascular cells as well as vascular cells. In fact, one study showed that retinal ganglion cells express CXCR4 (Chalasani et al., 2003).

CD34 is thought to be selectively expressed within the human and murine hematopoietic systems on stem and progenitor cells. In the current and previous studies, we find that CD34 is only expressed on endothelial cells of formed retinal vessels and not by retinal angioblasts (Chan-Ling et al., 2004). This was also true for CD31 (PECAM) in developing embryonic/fetal retina. In mouse embryo, Drake and Fleming (2000) reported that the immunolocalization of CD34 and CD31 was intense on aortic endothelial cells, moderate on primordial endothelial cells, and absent on angioblasts (TAL-1+). It is now widely accepted that endothelial cell (EC) progenitors (angioblasts) derived from adult blood play an important role in vascular repair. Blood-derived cells represent ∼10% of ECs in the neovasculature formed in response to surgical sponge implantation in mice (Crosby et al., 2000) and a similar fraction of blood-derived cells appears to be present in the neovasculature associated with angiogenesis in mice after induction of hindlimb ischemia (Asahara et al., 1999). Blood-derived cells also contribute to neovascularization in animal models of retinal and choroidal neovascularization (Grant et al., 2003; Sengupta et al., 2003). CD34 was thought to be a primary marker of the hemangioblast. However, Guo et al. (2003) have shown that FLK-1+/CD34−/CD31− mononuclear cells isolated from fetal bone marrow and cultured on ECM gel in the presence of endothelial cell growth factors, expressed CD34, proliferated, migrated, and formed a vascular plexus. They concluded that Flk1+/CD34−/CD31− cells may be more primitive than Flk1+/CD34+/CD31− cells in vascular development. Our studies (Chan-Ling et al., 2004) and unpublished results demonstrate that FLK1+/CD34−/CD31− cells in inner avascular fetal retina contribute to vasculogenesis, which suppports the theory of Guo et al. (2003).

There are 2 types of glial cells that migrate into the developing fetal human retina and contribute to the vascular complex, astrocytes and microglia (Provis, 2001). They have been studied extensively by others and have been shown to closely accompany formation of the primary vasculature. Chu et al. (2001) showed that astrocyte precursor cells (APCs), expressing Pax-2, migrate into the retina from the optic nerve head and differentiate into GFAP+ astrocytes (Chu et al., 2001). APCs proliferate and lie at or just peripheral to the edge of forming vasculature. In contrast to reports by others examining different species, our results in dog and in human (Taomoto et al., 2000; Chan-Ling et al., 2004), demonstrate that differentiated astrocytes (GFAP+) are rarely if ever in advance of retinal vasculatures forming by vasculogenesis. This is unlike what has been reported in rodents, where astrocytes are thought to form a template for vascular development (Fruttiger, 2002; Otani et al., 2002).

Finally, the CD39+/CXCR4+ angioblasts do not appear to be white blood cells. At all ages, the number of cells expressing CD45, leukocyte common antigen, was consistently lower than the number of angioblasts. Furthermore, MHCII expressing cells are very low in number or not present at all time points examined in this report (unpublished data).

In summary, our results demonstrate that a pool of CXCR4+ postmitotic progenitors are present early in the human embryonic retina and can give rise to CD39+ angioblasts, well before the onset of vasculogenesis. The presence of CXCR4 and CD39 in angioblasts and in the tips of forming blood vessels suggests that these two markers should be employed in future studies of angioblasts and vasculogenesis in other organ systems. The fact that the number of CXCR4+ cells is so great suggests that they are probably multipotent and that their fate is unlikely to be restricted to angioblasts. In fact, they may become neurons, astrocytes, microglia, or ganglion cells (Lazarini et al., 2003). Chalasani and associates have reported that CXCR4 was necessary for embryonic ganglion cell survival (Chalasani et al., 2003). It appears that the number of CXCR4+ declines as they differentiate into angioblasts and then endothelial cells, and possibly into neurons, astrocytes, microglia, or ganglion cells. These fates for CXCR4+ precursors in inner retina are logical since they are the cells of the inner neuroblastic layer in inner retina, which changes dramatically with development. Studies in our lab have examined stromal derived factor-1 (SDF-1), which binds exclusively to CXCR4, along with several other stem cell markers during embryonic/fetal retinal development and will be the subject of a separate report. From that study (unpublished data) in which SDF-1 is very prominent in developing inner retina, the trafficking of normal stem cells during development seems to be regulated to a significant degree by the SDF-1-CXCR4 axis. The high levels of CXCR4 expression in retinal angioblasts in vitro (Lutty et al., 2006) and in vivo suggests that its only ligand SDF-1 plays a central role in retinal vascular development.

Experimental Procedures

Age Determination and Preparation of Tissue

Thirty fetal human eyes from 18 fetuses, ranging in age from 6 to 23 WG, were used in this study. The ages and utilization of the tissue are given in Table 1. Tissues were provided by Advanced Bioscience Resources, Inc (Alameda, CA) after aspiration abortions in accordance with the guidelines set forth in the Declaration of Helsinki with the approval by the Joint Commission for Clinical Research at the Johns Hopkins University, School of Medicine. The age of each fetus was determined using last menstruation date and/or ultrasonography and fetal foot length as a reliable indicator of gestational age (Mhaskar et al., 1989). After enucleation, the eyes were immediately fixed at room temperature for 1 hr with 2% paraformaldehyde in 0.1 M sodium phosphate buffer pH 7.2 with 5% sucrose. The eyes then were washed in 0.1 M sodium phosphate buffer with 5% sucrose and were shipped overnight at 4°C. For older (>12 WG) cryopreserved eyes, the anterior segments were removed and the eye cups then were washed at room temperature in 0.1 M sodium phosphate buffer with increasing concentrations of sucrose: a 2:1, 1:1, and 1:2 mixture of 5% sucrose:20% sucrose (Lutty et al., 1993). The eyecups were held at room temperature for 2 hr in 20% sucrose in 0.1 M sodium phosphate buffer. The eyecups were dissected to make cryoblocks, usually quartered to contain the 4 arcades of retinal blood vessels in 16 WG and older eyes or the 2 vascular arcades (superior/inferior) and avascular regions in 12–14 WG younger eyes. For younger eyes (9 WG and younger), the whole eye was processed and frozen in toto. The tissue was placed in flat embedding molds with an embedding solution consisting of a 2:1 mixture of 20% sucrose in 0.1M sodium phosphate buffer:OCT compound (Tissue-Tek, Baxter Scientific, Columbia, MD) and infiltrated for 30 min at room temperature. The tissues were then frozen in isopentane cooled with dry ice. Cross-sections 8 μm thick were cut on a Reichert Jung (Deerfield, IL) Frigocut N cryostat at −25°C.

Table 1. Human Eyesa.

Age (WG) Cryo Wholemount
6 1
6.5 1
7 2
7.5 2 1 (CD39/CXCR4)
8 2 (CD39/CXCR4)
8.5 2 1 (CD39/CXCR4)
9 2
9.5 1 (CD39/CXCR4)
12 2 2 (CD34) (CD39)
14 1 2 (CD34) (ADP)
16 1 1 (ADP)
20 2
22 1 1 (CD39/CXCR4)
23 1 1 (ADP)
Open in a new tab
a

WG, weeks gestation; Cryo, cryopreserved; CD39/CXCR4, double labeled for CD39 and CXCR4 for confocal analysis; ADP, incubated for ADPase activity.

Alkaline Phosphatase Immunohistochemistry

Streptavidin alkaline phosphatase (APase) immunohistochemistry was performed on sections of cryopreserved tissue using a nitro blue tetrazolium (NBT) system recently developed by Bhutto et al. (2004). In brief, 8-μm-thick cryosections were permeabilized with absolute methanol and blocked with 2% normal goat serum in Tris-buffered saline (TBA; ph 7.4 with 1%BSA) and then an avidin-biotin complex (ABC) blocking kit (Vector Laboratories, Inc., Burlingame, CA). After washing in TBS, the sections were incubated overnight at 4°C with one of the following primary antibodies: mouse anti-CD39 (1:400, Chemicon, Temecula, CA); mouse anti-CD34 (1:800, Signet, Dedham, MA); mouse anti-CD31 (1:1,000, Dako, Carpinteria, CA); rabbit anti-CXCR4 (1:5000, Novus, Littleton, CO); mouse anti-CD45 (1:800, Chemicon); mouse anti-MHC class II (1:500, Dako); rabbit anti-Pax-2 (1:16,000, Covance/Babco, Berkeley, CA); rabbit anti-GFAP (1: 100,000, Dako); mouse anti-smooth muscle actin (1:16,000, Dako); rabbit anti-NG2 (1:2000, Chemicon); and mouse anti-Ki67 (1:1,000, Zymed, South San Francisco, CA). After they were washed in TBS, sections were incubated for 30 min at room temperature with the appropriate biotinylated secondary antibodies diluted 1:500 (Kirkegaard and Perry, Gaithersburg, MD). Finally, sections were incubated with streptavidin APase (1: 500; Kirkegaard and Perry) and the APase activity was developed with a 5-bromo-4-chloro-3-indoyl phosphate (BCIP)-NBT kit (Vector Laboratories, Inc.), yielding a blue immuno-reaction product.

Tissue Bleaching

For qualitative assessment of immunohistochemical staining at the level of Bruch's membrane-RPE complex, the removal of melanin pigment was desirable. Melanin in RPE was bleached using a slight modification of the technique recently developed by Bhutto et al. (2004). In brief, sections were fixed in 4% paraformaldehyde overnight at 4°C immediately after streptavidin APase immunohistochemistry. Slides were washed in distilled water at room temperature, immersed in a 0.05% potassium permanganate solution (Aldrich Chemical Co., Milwaukee, WI) for 25 min, and then rinsed in distilled water for 5 min. Sections were covered with 35% peracetic acid (FMC Corp., Philadelphia, PA) in a humidified container for 50 min at room temperature followed by washing in distilled water for 10 min twice. Finally, coverslips were mounted with Kaiser's aqueous mounting medium without counterstaining.

Double Labeling

Double label immunofluorescence was performed on cryopreserved tissue sections and a Zeiss 510 META confocal microscope (Carl Zeiss MicroImaging, Inc., Thornwood, NY) was used for analysis of the fluorescence (Wilmer Imaging Core Facility). Eight-micrometer-thick cryosections were permeabilized with absolute methanol and blocked with 2% normal goat serum in Tris-buffered saline (TBA; ph 7.4 with 1%BSA). After they were washed in TBS, the sections were incubated for 2 hr at room temperature by combining two of the following primary antibodies. For analysis of cell types that were proliferating, we used the following combinations: mouse anti-CD34 (1: 800, Signet, Dedham, MA)/rabbit anti-Ki67 (1:50, Zymed); mouse anti-CD39 (1:200, Chemicon)/rabbit anti-Ki67 (1:50); mouse anti-CD31 (1:1,000, Dako)/rabbit anti-Ki67 (1: 50); rabbit anti-Pax-2 (1:1,000, Covance/Babco, Berkeley, CA)/mouse anti-Ki67 (1:100); rabbit anti-GFAP (1:10,000, Dako)/mouse anti-Ki67 (1: 100); rabbit anti-CXCR4 (1:1000, Novus, Littleton, CO)/mouse anti-Ki67 (1:100). After washing in TBS, sections were incubated for 30 min at room temperature with the mixed appropriate secondary antibodies diluted 1:500 conjugated with Cy3 and diluted 1:100 conjugated with FITC. Finally, sections were mounted in DAPI counterstaining medium (Vector). Sections were observed with a Zeiss 510 META confocal microscope and the spectra of fluorescence were analysed. In addition to the proliferation series described above, we also double labeled sections with mouse anti-CD39 (1:200, Chemicon)/rabbit anti-CXCR4 (1:1000, Novus).

Whole Mounts

The tissues were fixed for 1 hr with 2% paraformaldehyde and then shipped overnight at 4°C in TBS. The anterior segments were removed from the eyes and the retinas separated from RPE. The retinas were then washed in TBS and permeabilized in absolute methanol for 15 min. Afterwards, the retinas were incubated with 1%Triton in TBS for 30 min. The retinas were blocked with 5% normal goat serum in TBS at 4°C overnight. After washing in TBS, retinas were incubated with mouse anti-CD39 (1:50; Chemicon) and rabbit anti-CXCR4 (1:250; Novus) for 72 hr at 4°C. After primary antibody labeling, the retinas were washed in TBS and then incubated with secondary antibodies for 48 hr at 4°C [goat anti-mouse secondary antibodies diluted 1:100 conjugated with Cy3 (Jackson ImmunoResearch) and goat-anti rabbit secondary antibody diluted 1:50 conjugated with FITC (Jackson ImmunoResearch)]. After secondary labeling, the retinas were washed in TBS and flattened. Immunofluorescence was then imaged with a Zeiss 510 META confocal microscope at 488 and 532 nm excitation (FITC and Cy3, respectively).

Confocal Microscopy

Tissue sections with fluorescent secondary antibodies and DAPI counterstain were examined with a Zeiss 510 META confocal microscope. Excitation wavelengths included 405, 488, and 532 nm for the DAPI, FITC, and Cy3, respectively. The tissues were analyzed with either 5, 10, 20, 40, or 63× objectives. Multispectral confocal microscopy was used as a validation tool for intracellular localization of antigens. Since multiple fluorescent colors were used to label various antigens and nuclei, it was necessary to use a multispectral confocal microscope that could separate the color optical overlaps in order to determine true co-localization and reveal autofluorescence. The method of spectral deconvolution was developed at JPL/Cal Tech (Pasadena, CA) and was implemented in a new generation of multispectral confocal microscope (Model 510 META, Zeiss, Inc.). The “emission fingerprinting” algorithm works by fitting the spectral components over low or non-overlapping portions of the combined spectrum of a multicolor image. The components are appropriately weighted so that the combination of color components matches the overall spectrum from the image pixel-by-pixel in each image plane.

Acknowledgments

This work has been supported in part by NIH-EY-09357 (G.L.), EY01765 (Wilmer), and a gift from the Himmelfarb Family Foundation in the name of Morton Goldberg, M.D. T. Hasegawa was a Bausch & Lomb Japan Research Fellow.

References

  1. Asahara T, Masuda H, Takahashi T, Kalka C, Pastore C, Silver M, Kearne M, Magner M, Isner JM. Bone marrow origin of endothelial progenitor cells responsible for postnatal vasculogenesis in physiological and pathological neovascularization. Circ Res. 1999;85:221–228. doi: 10.1161/01.res.85.3.221. [DOI] [PubMed] [Google Scholar]
  2. Ashton N. Retinal angiogenesis in the human embryo. Brit Med Bull. 1970;26:103–106. doi: 10.1093/oxfordjournals.bmb.a070758. [DOI] [PubMed] [Google Scholar]
  3. Beck L, D'amore PA. Vascular development: cellular and molecular regulation. FASEB J. 1997;11:365–373. [PubMed] [Google Scholar]
  4. Bhutto IA, Kim SY, Mcleod DS, Merges CA, Fukai N, Olsen BR, Lutty GA. Retinal and choroidal localization of collagen XVIII and the endostatin portion of collagen XVIII in aged human control and in age-related macular degeneration subjects. Invest Ophthalmol Vis Sci. 2004;45:1544–1552. doi: 10.1167/iovs.03-0862. [DOI] [PubMed] [Google Scholar]
  5. Chalasani S, Baribaud F, Coughlan C, Sunshine M, Lee V, Doms R, Littman DR, Raper JA. The chemokine stromal cell-derived factor-1 promotes the survival of embryonic retinal ganglion cells. J Neurosci. 2003;23:4601–4612. doi: 10.1523/JNEUROSCI.23-11-04601.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chan-Ling T, Mcleod DS, Hughes S, Baxter L, Chu Y, Hasegawa T, Lutty GA. Astrocyte-endothelial cell relationships during human retinal vascular development. Invest Ophthalmol Vis Sci. 2004;45:2020–2032. doi: 10.1167/iovs.03-1169. [DOI] [PubMed] [Google Scholar]
  7. Chu Y, Hughes S, Chan-Ling T. Differentiation and migration of astrocyte precursor cells and astrocytes in human fetal retina: relevance to optic nerve coloboma. FASEB J. 2001;15:2013–2015. doi: 10.1096/fj.00-0868fje. [DOI] [PubMed] [Google Scholar]
  8. Civin C, Strauss LC, Brovall C, Fackler MJ, Schwartz J, Shaper J. Antigenic analysis of hematopoiesis. III. A hematopoietic progenitor cell surface antigen defined by a monoclonal antibody raised against KG-1a cells. J Immunol. 1984;133:157–165. [PubMed] [Google Scholar]
  9. Crosby J, Kaminski W, Schatteman G, Martin P, Raines E, Seifert R, Bowen-Pope DF. Endothelial cells of hematopoietic origin make a significant contribution to adult blood vessel formation. Circ Res. 2000;87:728–730. doi: 10.1161/01.res.87.9.728. [DOI] [PubMed] [Google Scholar]
  10. Drake C, Fleming P. Vasculogenesis in the day 6.5 to 9.5 mouse embryo. Blood. 2000;95:1671–1679. [PubMed] [Google Scholar]
  11. Eltzschig H, Ibla J, Furuta GT, Leonard M, Jacobson K, Enjyoji K, Robson S, Colgan S. Coordinated adenine nucleotide phosphohydrolysis and nucleoside signaling in posthypoxic endothelium: role of ectonucleotidases and adenosine A2B receptors. J Exp Med. 2003;198:783–796. doi: 10.1084/jem.20030891. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fina L, Molgaard H, Robertson D, Bradley N, Monaghan P, Delia D, Sutherland D, Baker MA, Greaves MF. Expression of the CD34 gene in vascular endothelial cells. Blood. 1990;75:2417–2426. [PubMed] [Google Scholar]
  13. Fruttiger M. Development of the mouse retinal vasculature: angiogenesis versus vasculogenesis. Invest Ophthalmol Vis Sci. 2002;43:522–527. [PubMed] [Google Scholar]
  14. Goepfert C, Sundberg C, Sevigny J, Enjyoji K, Hoshi T, Csizmadia E, Robson S. Disordered cellular migration and angiogenesis in cd39-null mice. Circulation. 2001;104:3109–3115. doi: 10.1161/hc5001.100663. [DOI] [PubMed] [Google Scholar]
  15. Grant MB, Caballero S, Brown GA, Guthrie SM, Mames RN, Vaught T, Scott EW. The contribution of adult hematopoietic stem cells to retinal neovascularization. Adv Exp Med Biol. 2003;522:37–45. doi: 10.1007/978-1-4615-0169-5_5. [DOI] [PubMed] [Google Scholar]
  16. Guo H, Fang B, Liao L, Zhao Z, Liu J, Chen H, Hsu S, Cui Q, Zhao R. Hemangioblastic characteristics of fetal bone marrow-derived Flk1(+)CD31(−)CD34(−) cells. Exp Hematol. 2003;31:650–658. doi: 10.1016/s0301-472x(03)00087-0. [DOI] [PubMed] [Google Scholar]
  17. Hughes S, Yang H, Chan-Ling T. Vascularization of the human fetal retina: roles of vasculogenesis and angiogenesis. Invest Ophthalmol Vis Sci. 2000;41:1217–1228. [PubMed] [Google Scholar]
  18. Imhof BA, Dunon D. Leukocyte migration and adhesion. Adv Immunol. 1995;58:345–416. doi: 10.1016/s0065-2776(08)60623-9. [DOI] [PubMed] [Google Scholar]
  19. Juarez J, Bendall L, Bradstock K. Chemokines and their receptors as therapeutic targets: the role of the SDF-1/CXCR4 axis. Curr Pharm Des. 2004;10:1245– 1259. doi: 10.2174/1381612043452640. [DOI] [PubMed] [Google Scholar]
  20. Kaczmarek E, Koziak K, Sevigny J, Siegel JB, Anrather J, Beaudoin A, Bach F, Robson S. Identification and characterization of CD39/vascular ATP diphosphohydrolase. J Biol Chem. 1996;271:33116–33122. doi: 10.1074/jbc.271.51.33116. [DOI] [PubMed] [Google Scholar]
  21. Labastie M, Cortes F, Romeo PH, Dulac C, Peault B. Molecular identity of hematopoietic precursor cells emerging in the human embryo. Blood. 1998;92:3624– 3635. [PubMed] [Google Scholar]
  22. Lazarini F, Tham TO, Casanova P, Arenzana-Seisdedos F, Dubois-Dalcq M. Role of the a-chemokine stromal cell-derived factor (SDF-1) in the developing and mature central nervous system. Glia. 2003;42:139–148. doi: 10.1002/glia.10139. [DOI] [PubMed] [Google Scholar]
  23. Lutty GA, McLeod DS. A new technique for visualization of the human retinal vasculature. Arch Ophthalmol. 1992;110:267–276. doi: 10.1001/archopht.1992.01080140123039. [DOI] [PubMed] [Google Scholar]
  24. Lutty GA, McLeod DS. Retinal vascular development and oxygen-induced retinopathy: a role for adenosine. Prog Ret Eye Res. 2003;22:95–111. doi: 10.1016/s1350-9462(02)00058-7. [DOI] [PubMed] [Google Scholar]
  25. Lutty GA, McLeod DS. Phosphatase enzyme histochemistry for studying vascular hierarchy, pathology, and endothelial cell dysfunction in retina and choroid. Vision Res. 2005;45:3504–3511. doi: 10.1016/j.visres.2005.08.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Lutty GA, Merges C, Threlkeld AB, Crone S, McLeod DS. Heterogeneity in localization of isoforms of TGF-b in human retina, vitreous, and choroid. Invest Ophthalmol Vis Sci. 1993;34:477–487. [PubMed] [Google Scholar]
  27. Lutty G, McLeod D, Pachnis A, Costantini F, Fabry M, Nagel R. Retinal and choroidal neovascularization in a transgenic mouse model of sickle cell disease. Am J Pathol. 1994;145:490–497. [PMC free article] [PubMed] [Google Scholar]
  28. Lutty GA, Phelan A, McLeod DS, Fabry ME, Nagel RL. A rat model for sickle cell-mediated vaso-occlusion in retina. Microvascular Res. 1996;52:270–280. doi: 10.1006/mvre.1996.0064. [DOI] [PubMed] [Google Scholar]
  29. Lutty GA, Merges C, Grebe R, Prow T, McLeod DS. Canine retinal angioblasts are multipotent. Exp Eye Res. 2006;83:183–193. doi: 10.1016/j.exer.2005.09.025. [DOI] [PubMed] [Google Scholar]
  30. Mann IC. The development of the human eye. Cambridge: University Press; 1928. [Google Scholar]
  31. McLeod DS, Lutty GA, Wajer SD, Flower RW. Visualization of a developing vasculature. Microvasc Res. 1987;33:257–269. doi: 10.1016/0026-2862(87)90021-5. [DOI] [PubMed] [Google Scholar]
  32. McLeod D, Goldberg M, Lutty G. Dual perspective analysis of vascular formations in sickle cell retinopathy. Arch Ophthalmol. 1993;111:1234–1245. doi: 10.1001/archopht.1993.01090090086026. [DOI] [PubMed] [Google Scholar]
  33. McLeod DS, Brownstein R, Lutty GA. Vaso-obliteration in the canine model of oxygen-induced retinopathy. Invest Ophthalmol Vis Sci. 1996a;37:300–311. [PubMed] [Google Scholar]
  34. McLeod D, Crone SN, Lutty GA. Vasoproliferation in the neonatal dog model of oxygen-induced retinopathy. Invest Ophthalmol Vis Sci. 1996b;37:1322–1333. [PubMed] [Google Scholar]
  35. McLeod, DS, Merges C, Fukushima A, Goldberg MF, Lutty GA. Histopathological features of neovascularization in sickle cell retinopathy. Am J Ophthalmol. 1997;124:473–487. doi: 10.1016/s0002-9394(14)70862-1. [DOI] [PubMed] [Google Scholar]
  36. Mhaskar R, Agarwal N, Takkar D, Buck-shee K, Deorari A. Fetal foot length: a new parameter for assessment of gestational age. Int J Gynaecol Obstet. 1989;29:35–38. doi: 10.1016/0020-7292(89)90126-4. [DOI] [PubMed] [Google Scholar]
  37. Newman PJ. The biology of PE-CAM-1. J Clin Invest. 1997;100:S25–29. [PubMed] [Google Scholar]
  38. Otani A, Kinder K, Ewalt K, Otero FJ, Schimmel P, Friedlander M. Bone marrow-derived stem cells target retinal astrocytes and can promote or inhibit retinal angiogenesis. Nat Med. 2002;8:1004– 1010. doi: 10.1038/nm744. [DOI] [PubMed] [Google Scholar]
  39. Peichev M, Naiyer AJ, Pereira D, Zhu Z, Lane WJ, Williams M, Oz MC, Hicklin DJ, Witte L, Moore MA, Rafii S. Expression of VEGFR-2 and AC133 by circulating human CD34(+) cells identifies a population of functional endothelial precursors. Blood. 2000;95:952–958. [PubMed] [Google Scholar]
  40. Provis JM. Development of the primate retina. Prog Retin Eye Res. 2001;20:799– 821. doi: 10.1016/s1350-9462(01)00012-x. [DOI] [PubMed] [Google Scholar]
  41. Rhodes R. A light microscopic study of the developing human neural retina. Am J Anat. 1979;154:195–209. doi: 10.1002/aja.1001540206. [DOI] [PubMed] [Google Scholar]
  42. Risau W. Mechanisms of angiogenesis. Nature. 1997;386:671–674. doi: 10.1038/386671a0. [DOI] [PubMed] [Google Scholar]
  43. Risau W. Development and differentiation of endothelium. Kidney Int. 1998;54(Suppl):S3–S6. doi: 10.1046/j.1523-1755.1998.06701.x. [DOI] [PubMed] [Google Scholar]
  44. Risau W, Flamme I. Vasculogenesis. Ann Rev Cell Dev Biol. 1995;11:73–91. doi: 10.1146/annurev.cb.11.110195.000445. [DOI] [PubMed] [Google Scholar]
  45. Sandercoe TM, Madigan MC, Billson FA, Penfold PL, Provis JM. Astrocyte proliferation during development of the human retinal vasculature. Exp Eye Res. 1999;69:511–523. doi: 10.1006/exer.1999.0730. [DOI] [PubMed] [Google Scholar]
  46. Sengupta N, Caballero S, Mames RN, Butler JM, Scott EW, Grant MB. The role of adult bone marrow-derived stem cells in choroidal neovascularization. Invest Ophthalmol Vis Sci. 2003;44:4908–4913. doi: 10.1167/iovs.03-0342. [DOI] [PubMed] [Google Scholar]
  47. Taomoto M, McLeod DS, Merges C, Lutty GA. Localization of adenosine A2a receptor in retinal development and oxygen-induced retinopathy. Invest Ophthalmol Vis Sci. 2000;41:230–243. [PubMed] [Google Scholar]
  48. Tavian M, Coulombel L, Luton D, Clemente H, Dieterlen-Lievre F, Peault B. Aorta-associated CD34+ hematopoietic cells in the early human embryo. Blood. 1996;87:67–72. [PubMed] [Google Scholar]
  49. Yong K, Watts M, Shaun Thomas N, Sullivan A, Ings S, Linch DC. Transmigration of CD34+ cells across specialized and nonspecialized endothelium requires prior activation by growth factors and is mediated by PECAM-1 (CD31) Blood. 1998;91:1196–1205. [PubMed] [Google Scholar]

RESOURCES