Molecular Vision 2009; 15:45-59 <>
Received 24 October 2008 | Accepted 22 December 2008 | Published 12 January 2009

Differential expression of anti-angiogenic factors and guidance genes in the developing macula

Peter Kozulin,1 Riccardo Natoli,1 Keely M. Bumsted O’Brien,1 Michele C. Madigan,2 Jan M. Provis1,3

1ARC Centre of Excellence in Vision Science and School of Biology, The Australian National University, Canberra, Australia; 2School of Optometry and Vision Science, The University of New South Wales, Kensington, and Save Sight Institute, The University of Sydney, Sydney, Australia; 3Australian National University Medical School, The Australian National University, Canberra, Australia

Correspondence to: Jan Provis, School of Biology, Building 46, Biology Place, The Australian National University, Canberra ACT 0200, Australia; Phone: +61 2 6125 4242; FAX: +61 2 6125 8680; email:


Purpose: The primate retina contains a specialized, cone-rich macula, which mediates high acuity and color vision. The spatial resolution provided by the neural retina at the macula is optimized by stereotyped retinal blood vessel and ganglion cell axon patterning, which radiate away from the macula and reduce shadowing of macular photoreceptors. However, the genes that mediate these specializations, and the reasons for the vulnerability of the macula to degenerative disease, remain obscure. The aim of this study was to identify novel genes that may influence retinal vascular patterning and definition of the foveal avascular area.

Methods: We used RNA from human fetal retinas at 19–20 weeks of gestation (WG; n=4) to measure differential gene expression in the macula, a region nasal to disc (nasal) and in the surrounding retina (surround) by hybridization to 12 GeneChip® microarrays (HG-U133 Plus 2.0). The raw data was subjected to quality control assessment and preprocessing, using GC-RMA. We then used ANOVA analysis (Partek® Genomic Suite™ 6.3) and clustering (DAVID website) to identify the most highly represented genes clustered according to “biological process.” The neural retina is fully differentiated at the macula at 19–20 WG, while neuronal progenitor cells are present throughout the rest of the retina. We therefore excluded genes associated with the cell cycle, and markers of differentiated neurons, from further analyses. Significantly regulated genes (p<0.01) were then identified in a second round of clustering according to molecular/reaction (KEGG) pathway. Genes of interest were verified by quantitative PCR (QRT–PCR), and 2 genes were localized by in situ hybridization.

Results: We generated two lists of differentially regulated genes: “macula versus surround” and “macula versus nasal.” KEGG pathway clustering of the filtered gene lists identified 25 axon guidance-related genes that are differentially regulated in the macula. Furthermore, we found significant upregulation of three anti-angiogenic factors in the macula: pigment epithelium derived factor (PEDF), natriuretic peptide precurusor B (NPPB), and collagen type IVα2. Differential expression of several members of the ephrin and semaphorin axon guidance gene families, PEDF, and NPPB was verified by QRT–PCR. Localization of PEDF and Eph-A6 mRNAs in sections of macaque retina shows expression of both genes concentrates in the ganglion cell layer (GCL) at the developing fovea, consistent with an involvement in definition of the foveal avascular area.

Conclusions: Because the axons of macular ganglion cells exit the retina from around 8 WG, we suggest that the axon guidance genes highly expressed at the macula at 19–20 WG are also involved in vascular patterning, along with PEDF and NPPB. Localization of both PEDF and Eph-A6 mRNAs to the GCL of the developing fovea supports this idea. It is possible that specialization of the macular vessels, including definition of the foveal avascular area, is mediated by processes that piggyback on axon guidance mechanisms in effect earlier in development. These findings may be useful to understand the vulnerability of the macula to degeneration and to develop new therapeutic strategies to inhibit neovascularization.


A unique adaptation of the central retina in monkeys and humans is the distinctive macula, at the center of which is the highly specialized fovea centralis (fovea; Figure 1). Within the macula the density of all neural elements is elevated and there are no large retinal vessels. At the fovea there are no rods, but there is a peak spatial density of cone photoreceptors, and a prevalence of “midget” circuitry that mediates high acuity and color vision [1]. Several retinal diseases specifically affect the macula, including Best's disease, macular telangiectasia, Stargardt's macular dystrophy, and age-related macular degeneration (AMD). It is not known why the macula is the focus of these disorders. AMD is the most common cause of untreatable blindness, resulting from degeneration of macular photoreceptors, with or without neovascularization [2]. While recent gene association studies attribute high risk for AMD to a relatively small number of SNPs [3-6], why the disease specifically affects the macula remains poorly understood. Because of the vulnerability of the macula, and the high impact consequences of macular disease, there has been considerable interest in identifying genes differentially expressed in the macula and foveal region [7-10]. To our knowledge, however, no studies have investigated differential gene expression in the macula during development, that is, during the period when genes that are preferentially associated with determination of macular characteristics are most likely to be expressed.

In all species examined, the fovea is found within an area devoid of retinal vessels - the foveal avascular area (Figure 1) [11]. In humans retinal vessels enter the developing human retina in the 14th week of fetal life (14 weeks of gestation; WG) and grow slowly toward the developing macula, compared with the rest of the retina [12]. In normal development, the foveal region is never vascularized [13], and absence of a foveal avascular region is associated with foveal hypoplasia [14,15]. During fetal life, the macula is the most developmentally advanced region of the neural retina, and the principal angiogenic factor, vascular endothelial growth factor (VEGF), is highly expressed by ganglion cells in the macula [16]. Despite this, blood vessels growing into the macular region proceed slowly [12], have lower rates of proliferation [17], and appear to face a molecular barrier that prevents formation of vessels at the incipient fovea [18,19]. Developmental studies of macaque and marmoset retinas also indicate that the fovea forms only after the foveal avascular area has been defined [18,20].

The photoreceptor layer is normally avascular throughout life, and neovascularization of the outer retina is associated with vision loss, as in exudative AMD. It is thought that a critical balance between the pro-angiogenic factor VEGF and the anti-angiogenic pigment epithelial derived factor (PEDF) is essential for preserving the avascular environment of the outer retina [21]. Both of these factors are expressed by the retinal pigmented epithelium. However, to date there have been no reports of differential expression of PEDF or other anti-angiogenic factors that may be involved in definition of the foveal avascular area, even though data from several gene expression studies of adult retina have been reported [9,10]. Furthermore, because guidance factors commonly define “no-go” zones for axons, and members of the ephrin, netrin, slit, and semaphorin families have roles in both axon and vascular guidance [22], we hypothesized that a gradient of guidance molecules centered on the incipient fovea might also participate in definition of the foveal avascular area. Our aim, therefore, was to determine if genes associated with anti-angiogenic regulation as well as axon and vascular guidance are differentially expressed in the macula during development.


Human retinas

Human retinas (n=18) were obtained at surgery to terminate pregnancy from fetuses between 17 and 20 WG, with informed maternal consent and ethical approval from Human Ethics Committees of The University of Sydney and The Australian National University. Gestational age was determined by ultrasound before surgery and confirmed by postmortem measurements of foot length.

RNA extraction

For hybridization to microarrays (4 retinas), eyes were removed and stored in separate solution-free containers, on ice, prior to dissection. RNA was extracted from these retinas 90-120 min post mortem. For QRT-PCR, whole eyes were placed in TRIzol Reagent (Invitrogen, Carlsbad, CA) and frozen at -80 °C for subsequent RNA extraction, up to one month post mortem. The anterior segment and vitreous were dissected away, and the eye cup was flattened by making radial cuts through the sclera and retina. A 5 mm trephine was used to take two biopsies of each retina: one from the posterior pole (“macula” sample), and a second from the corresponding region immediately nasal to the optic disc (“nasal” sample). The remaining retina was used as a third sample region (“surround” sample). The biopsied regions are illustrated in Figure 1C. Each of the three retinal biopsies (macula, nasal, and surround) was homogenized then combined with TRIzol® Reagent (Invitrogen, Carlsbad, CA) and chloroform to separate the RNA from other cell components. The RNA was then purified and isolated using RNAqueous®-Micro RNA isolation kit (Ambion, Inc., Austin, TX). RNA concentration was determined using a NanoDrop® ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE). The RNA Integrity Number (RIN – determined by the Agilent Bioanalyzer 2100 eukaryotic total RNA Nano assay, Agilent Technologies, Inc., Santa Clara, CA) was used as a measure of the quality of the RNA [23]. In some cases where genomic DNA contamination was a concern, the RNA was DNase treated [24]. This treatment did not affect the RIN.


Three sample regions (macula, nasal, and surround) from each of four retinas, three at 19 WG and one at 20 WG were used for microarray analysis. Double-stranded cDNA was generated from each of the twelve total-RNA samples and transcribed to obtain biotin-labeled cRNA. The target cRNAs were fragmented, combined with hybridization and spike controls, then hybridized to the twelve HG-U133 Plus 2.0 GeneChip® microarrays (Affymetrix, Inc., Santa Clara, CA) for 16 h. Because of the limited availability of suitable donor retinas, six cDNA samples (from two retinas) were hybridized to microarrays on two separate occasions, that is, the hybridizations were performed in two batches. Following hybridization, the arrays were washed and stained with streptavidin phycoerythrin conjugate using an automated GeneChip® Fluidics Station 450, and then scanned with a GeneChip® Scanner 3000 using a 570 nm excitation wavelength laser. Target sample preparation and microarray hybridization and scanning were performed at the Biomolecular Resource Facility, John Curtin School of Medical Research, The Australian National University (Canberra, Australia).

To allow correct alignment of the software analysis grid, a control oligo (B2 oligo) was hybridized and used to indicate the boundaries of the probe area. Other quality control metrics used include average background and noise (Raw Q) of the measured probe cell intensities, ratio of expression of the 3′ probe set to 5′ probe set of the housekeeping control gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and the percentage of all probe sets on the array detected as present. Poly(A) sense RNA labeling controls and antisense biotinylated cRNA hybridization controls (together with the B2 oligo) were also spiked into the total RNA and biotinylated cRNA cocktails, respectively, at known concentrations before array hybridization (Affymetrix, Inc., GeneChip Expression Analysis technical manual; 2003). The poly(A) RNA controls were generated from four cloned B. subtilis genes (lys, phe, thr, and dap), and the hybridization controls transcribed from three E. coli genes (bioB, bioC, and bioD) and one P1 bacteriophage recombinase gene (cre).

The perfect match (PM) and mismatch (MM) probe intensities were used to generate a detection call (present, marginal, or absent) for each probe set, to assess the reliability of detection of a transcript. Detection calls were generated by the microarray suite version 5 (MAS5) algorithm. Microarray quality controls were assessed using GeneChip® Operating Software (Affymetrix).

Data analysis - preprocessing

Affymetrix CEL files were imported into Partek® Genomics Suite™ 6.3 (Partek Inc., St. Louis, MO) using the default Partek normalization parameters. Probe-level data was pre-processed, including background correction, normalization, and summarization [25], using robust multi-array average (RMA) analysis adjusted for probe sequence and GC content (GC-RMA). GC-RMA adjusts for background noise on each array using only the PM probe intensities; subsequent data normalization was performed across all arrays using quantile normalization [26,27]. The background-adjusted, normalized PM values were then compiled, or summarized, within each probe set, using the median polish technique, to generate a single measure of expression [27]. These expression measures were then log transformed, base 2.

Data analysis - differential expression analysis

Differential expression analysis was performed using a two-way ANOVA in which the variables were sample region (macula, nasal, and surround) and hybridization batch. No significant batch effect was evident in the data. Data from the four macula arrays were grouped and compared (using a linear contrast) with either the four nasal or surround arrays as baselines. The lists of differentially expressed genes from the macula versus surround and macula versus nasal data were clustered according to functional roles or biological process as specified by the Gene Ontology Consortium [28]. This was achieved using the functional annotation tool on the Database for Annotation, Visualization, and Integrated Discovery (DAVID) website [29] with an Expression Analysis Systematic Explorer (EASE) score threshold (a modified Fisher Exact p-value) of 0.1.

The retina differentiates in a gradient which progresses from the macula to the periphery. At 19–20 WG (the ages of our samples), the majority of cells in the macula have exited the cell cycle and differentiated as neurons, while the majority of cells in non-macular locations are still proliferating [30]. Because of these significant differences in cell function between macular and non-macular locations, we excluded genes that clustered under the general heading of “biological processes” from further differential analysis. The remaining list of genes was then subcategorized according to the molecular/reaction pathways described by the Kyoto Encyclopedia of Genes and Genomes (KEGG) using the DAVID web tool and 0.1 EASE score threshold.

Quantitative real time polymerase chain reaction

cDNA from human retinas aged between 17 and 20 WG were used for evaluative quantitative real-time polymerase chain reaction (QRT–PCR) analyses. First-strand cDNA synthesis was performed in a 20 μl reaction mixture using 1 μg RNA, 500 ng oligo (dT)18 primer, and 200 U SuperScript™ III reverse transcriptase (Invitrogen). Amplification of Eph-A6 and ephrin A1–5 was measured using SYBR® GreenER™ (Invitrogen). All other genes were amplified using commercially available TaqMan® probes (Applied Biosystems, Melbourne, Australia). Primers were designed within the coding domain sequence and across an intron using reference sequences (RefSeq) derived from the NCBI Entrez Gene database and the OligoPerfect™ Designer web-based program (Invitrogen). Primer sequences and TaqMan probe IDs are provided in Table 1 and Table 2, respectively. Fluorescence was measured on a FAM 510 nm detection channel using a commercial device (Rotor-Gene RG-3000; Corbett Research, Sydney, Australia) and normalized relative to a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) reference gene. TaqMan amplification was performed in duplicate, and SYBR GreenER PCR runs in triplicate, using cDNA from three different specimens (total of 6 and 9 reactions per gene, respectively). Amplification specificity was assessed using either melt curve analysis (SYBR GreenER) or gel electrophoresis (TaqMan). The results were analyzed using the relative quantification method developed by Pfaffl [31].

Macaque retinas and in situ hybridization

For localization studies we used histological sections of retinas from eight macaques of known gestational ages: fetal day (Fd) 80, Fd 115, and Fd 150. The sections were obtained by Prof. A.E. Hendrickson from the breeding colony of the Primate Center at Bogor Agricultural University, Indonesia, using protocols approved by the University of Washington (Seattle, WA) Animal Care Committee and in compliance with guidelines of the Association for Research in Vision and Ophthalmology. Fetuses were delivered by aseptic caesarean section, euthanized by intraperitoneal overdose of barbiturate, and then enucleated. Eyes were immediately injected with methyl Carnoy’s fixative (methyl alcohol: acetic acid: chloroform, 6:3:1), then fixed whole by immersion for 1–2 h, as described previously [32]. Sections were paraffin-embedded then cut at 8 μm. Only sections passing through the optic disc and foveal region were used for quantitative analysis in this study.

Eph-A6 and PEDF were cloned from PCR products using total RNA of human fetal retinas, the pGEM®-T DNA vector system and JM109 competent cells (#A3600; Promega, Madison, WI). A DIG RNA Labeling Kit SP6/T7 (#1175025; Roche, Basel, Switzerland) was used to transcribe linearized plasmid and generate DIG-labeled antisense and sense probes. The Eph-A6 and PEDF riboprobes were hybridized overnight at 55 °C and 58 °C, respectively, and then washed in saline sodium citrate (pH 7.4) at 65 °C and 58 °C, respectively. Paraffin sections of fetal macaque retina were processed as described previously [16].


The RNA extracted from human fetal retinas and hybridized to the arrays was of high quality and purity (Table 3). Microarray quality controls were within manufacturer recommendations, and control values were comparable between arrays (Table 3). In addition, on all arrays the average signal values of the poly(A) RNA and hybridization controls increased relative to the order of spike concentrations (lys, phe, thr, dap and bioB, bioC, bioD, cre, respectively) indicating efficient labeling and hybridization of all samples. The full set of microarray data has been deposited with the NCBI Gene Expression Omnibus repository under accession number GSE12621.

We generated two lists of differentially expressed genes: macula versus surround; and macula versus nasal. We selected a p-value cut-off of <0.01 which gave a false discovery rate (step up method) of 14%. Using DAVID, we found the differentially expressed genes were clustered according to their functional roles, or biological process, generating over 200 gene ontology clusters. Small clusters were consolidated under their parent categories to generate a simplified list (Table 4) that shows the 15 most highly represented biological processes in the microarray data sets. The largest cluster identified includes genes involved with cell metabolism (macula compared to surround: 1,520 genes; macula compared to nasal: 1,059 genes). Differential expression of genes in this cluster, including for example medium- and long-wavelength sensitive cone opsins, glucose phosphate dehydrogenase, and retinol dehydrogenase, is most likely due to the fact that cells in the macula exit the cell cycle and become mature neurons ahead of the rest of the retina. Other categories in which differential gene regulation was expected (because of the more highly differentiated state of macula versus non-macular neurons at 19–20 WG) and confirmed, include neurotransmitter transport (GO:0015813, GO:0006836), secretion (GO:0007269), and signaling (GO:0007215); cell cycle (GO:0007049), cell proliferation (GO:0008283), and division (GO:0051301); and response to light stimulus (GO:0009416).

In our further analyses we restricted the categories of gene ontology terms to exclude differentially regulated genes associated with the precocious development of the macula. Of the genes reported in Table 4, 52% of the “macula compared to surround” list of differentially expressed probe sets, and 43% of the “macula compared to nasal” list were thus eliminated, leaving a total of 1,788 differentially expressed genes in the “macula versus surround” list and 2,038 in the and “macula versus nasal” list. Further clustering of the restricted lists revealed genes that clustered into five common and highly represented KEGG pathways between the two lists: axon guidance (hsa04360), regulation of actin cytoskeleton (hsa04810), extracellular matrix-receptor interaction (hsa04512), focal adhesion (hsa04510), and ABC transporters-general (hsa02010). The axon guidance KEGG pathway was the most highly represented in the “macula cf surround” list. Table 5 lists 25 genes with previously identified roles in axon guidance common to both lists. Seven of the 35 probe sets listed interrogate members of the semaphorin/plexin/neuropilin gene family and six interrogate members of the Eph/ephrin gene family.

To identify levels of expression of genes involved in the regulation of angiogenesis, we assembled the gene expression data for the 17 negative regulators of angiogenesis identified in the Gene Ontology Consortium, and for isoforms of the major retinal angiogenic factor, VEGF (Appendix 1). There was no significant differential regulation of any of the isoforms of VEGF detectable by the microarrays (Appendix 1). Of 17 negative regulators of angiogenesis, three—PEDF, natriuretic peptide precursor B (NPPB), and collagen type IVα2—showed statistically significant fold-changes (p<0.01) in one or both sets of data (Table 6).

To test the veracity of the microarray data, we evaluated the expression of eleven genes from Table 5 and Table 6 and five additional genes of interest (including binding partners of significantly regulated genes, closely related genes, and significant angiogenic factors) using cDNA from retinas aged between 17 and 20 WG by QRT–PCR. Figure 2 shows the mean relative level of expression±standard error determined by QRT–PCR (red columns) compared with differential expression data from the microarray data (blue columns). The results show strong consensus between the microarray and QRT–PCR data for virtually all genes. The graphs confirm downregulation of neuropilin 1 (NRP1) and plexin C1 (PLXNC1) in the macula, and upregulation of the other selected genes in the macula. Comparison of fold-changes in “macula compared to surround” and “macula compared to nasal” (Figure 2A,B) revealed no significant difference between the two data sets. In subsequent QRT–PCR experiments we compared expression of ephrin A1-A5 and Eph-A6 (Figure 2C), and four regulators of angiogenesis (Figure 2D) in cDNA from macula and nasal locations. The results indicated strong upregulation of Eph-A6 in the macula, and downregulation of its 2 ligands, ephrin-A1 and -A4, as well as ephrin-A2 and -A3. Ephrin-A5 did not amplify reliably (data not shown). We found no differential regulation of VEGFA by QRT–PCR, consistent with the microarray data. We detected evidence of borderline downregulation of fibroblast growth factor 2 (FGF2) in the macula, consistent with a prior report [32]. Significantly, QRT–PCR confirmed strong upregulation of two negative regulators of angiogenesis: NPPB [33] and PEDF [34] (Figure 2D).

We selected one axon guidance gene, Eph-A6, and one negative regulator of angiogenesis, PEDF, to examine their retinal distributions by in situ hybridization. Figure 3A shows a macaque retina at Fd 115, equivalent to around 25 WG in the human, hybridized to show the retinal distribution of Eph-A6 mRNA. In this retina, and others from macaques aged Fd 110 to postnatal 3 months (data not shown), Eph-A6 mRNA expression was high in the macula, reaching a peak at the developing fovea, and was lower in the periphery [35]. We also detected a graded expression of Eph-A6 across the thickness of the ganglion cell layer (GCL) that is high in the inner GCL and low in the outer GCL at foveal and parafoveal locations, but not peripheral locations (Figure 3). In situ hybridization for PEDF in developing macaque retinas (Figure 4A-D) shows PEDF mRNA in the GCL and RPE, consistent with previous reports [36], although we found low to very low PEDF expression in the developing photoreceptors of fetal macaque retinas (Figure 4A-D). PEDF mRNA expression was low in the RPE at Fd 80 (about 17 WG in humans), but higher at Fd 150 (about 32 WG in humans; Figure 4A,B compared to Figure 4C,D). We detected PEDF expression at higher levels in the GCL of central retina, compared with non-central retina, at both Fd 80 and Fd 150. This gradient of expression of PEDF in the GCL is consistent with the microarray and QRT–PCR data.


To our knowledge, this is the first investigation into the differential gene expression in the developing macula. Previous differential expression studies have used adult tissue [37] or have not specifically investigated the developing macula [38].

Differential expression of anti-angiogenic factors

This is the first report of upregulation of anti-angiogenic factors in the primate macula: the widely recognized inhibitor PEDF, and a novel anti-angiogenic regulator for the retina, NPPB. The apparent downregulation of collagen type IVα2 in the macula is most likely due to the presence of collagen-invested retinal vessels in the non-macular samples. Our findings are consistent with a role for both PEDF and NPPB (also known as brain natriuretic peptide) in the definition of the avascular area in the central macula, which, at a later stage of development, will include the fovea centralis.

PEDF is a neurotrophin [39] and has potent anti-angiogenic effects in human [36] and other mammalian retinas [40,41]. The anti-angiogenic effects of PEDF occur through stimulation of the Fas/Fas ligand-mediated apoptotic pathway which targets the endothelial cells of immature vessels, and may also downregulate the caspase-8 inhibitor, c-FLIP [42]. Both of these effects reduce the survival of new endothelial cells, even in the presence of pro-angiogenic factors. VEGF-induced angiogenesis can be directly inhibited by PEDF via the intracellular proteolysis of VEGFR-1 or inhibition of its phosphorylation [43].

Less is known about NPPB in the retina. NPPB is a cardiac hormone and influences several systemic functions such as body fluid homeostasis, blood pressure, cardiac function, bone growth, and vascular tone [44]. NPPB has been amplified from adult human retina [33] and the protein localized to ganglion, Müller [33,45], and amacrine cells [46] in humans and rats. Natriuretic peptides can affect VEGF-induced angiogenesis by inhibiting certain signaling molecules such as ERK, JNK, and p38 members of the MAP kinase family through stimulation of a cyclic GMP signaling cascade (natriuretic peptide receptors NPR-A and -B) or through a less well understood mechanism involving NPR-C [44,47-49].

Type IV collagen is an important structural component of basement membrane and is required for blood vessel formation [50]. The 24 kDa C-terminal domain of α2 chains of type IV collagen, also known as canstatin, inhibits endothelial cell proliferation and migration, and induces other anti-angiogenic mechanisms including Fas ligand expression (reviewed in [51]). One would expect type IV collagen to be restricted to the more mature retinal vessels, which at 19–20 WG, are mainly in nasal and temporal superior/inferior retina; very few vessels would have been biopsied in our macular samples.

Differential expression of guidance factors

We have also identified 25 axon guidance genes differentially expressed in the macula, including 12 members of the ephrin, semaphorin, slit and netrin families, which are also regulators of vascular development [52]. Significantly, we detect this differential regulation of guidance genes at the macula 8–12 weeks after macular ganglion cells differentiate and send out axons, and 3–4 weeks after the majority of retinal axons have exited the retina [30]. Retinal neurons are generated in a stereotyped sequence [53] and along a centro-peripheral gradient [30,54], such that ganglion cells and cones of the incipient fovea are among the first neurons generated in the primate retina. Ganglion cell axons are observed in the optic stalk from around 8 WG in humans [55] (reviewed in [56]), reach a peak number at 16–17 WG, and stabilize at adult values between 25 and 30 WG, approximately [57]. Although for the most part, axon guidance within the retina and visual targets has not been studied in primate retina (see, however, Lambot et al., [58]), we know from studies of other species that a range of axon guidance factors are expressed in the retina during the period of axon outgrowth, to guide axons to the optic disc [59,60], and regulate targeting of their axon terminals by interaction with their binding partners in the primary visual nuclei [61-65]. Because axons of macular ganglion cells are among the first to enter the optic nerve—estimated between 8 and 14 WG—we interpret the sustained expression of guidance factors in the macula at 19–20 WG, as detected in the present study, to be primarily related to vascular patterning, rather than axon guidance.

Many genes originally identified by their involvement in axon patterning are also implicated in vascular patterning. The semaphorin-plexin family of genes share with VEGFA the capacity to bind neuropilin 1, expressed by both blood vessels and axons. Class 3 semaphorins are also known to have a repellent effect during vascular morphogenesis via interactions with integrins [66]. Eph receptors and their ephrin ligands have key roles in axon guidance, provide guidance cues for endothelial cells during development, are involved in assembly and maintenance of vascular networks, and arteriovenous differentiation [67]. Netrin is a potent vascular mitogen [22] and has a role in repelling developing vessels via interactions with the UNC5 receptor [68], while Slit2 is implicated in endothelial cell migration [52]. Of particular interest are the repellent effects of netrin-UNC5 interactions [68] and Eph-ephrin signaling [69] on developing vessels as well as axons during development. We propose that a graded expression of genes involved in repellent signaling that is centered on the fovea during development - similar to the one we report here for Eph-A6 (Figure 3) - retards the growth of vessels into the central region of the retina, and contributes to definition of the foveal avascular area (in preparation).


The spatial resolving power of the fovea is directly attributable to cone inner segment spacing which reaches a peak of 100,000 to 300,000 mm−2 [70] at the geometric center of the avascular area, and corresponds with the central fovea (foveola) in adult retinas. Studies using adaptive optics show that while most individuals do not use the area of peak cone density for fixation, the mean displacement of the fixation point from the peak cone density (50 µm) [71] places it comfortably within the average-sized avascular area (500–700 µm). Shadowing of the photoreceptor mosaic causes angioscotomas in the visual cortex [72]. Thus, reduction in the diameter of macular vessels, and an absence of vessels overlying the region of peak cone density at the fovea, might be regarded as an adaptive advantage that favors optimal spatial acuity. It is possible that the mechanisms that mediate specialization of the macular vessels, and which define the foveal avascular area, “piggyback” on axon guidance mechanisms in effect earlier in development. One hypothesis is that the guidance factors that repel ganglion cell axons away from central retina into the arcuate axon bundles and optic nerve head also guide vessels out of the optic nerve head and repel them from the foveal region, at a slightly later phase of development.

A series of findings suggest that definition of an avascular area within the macula is required for subsequent formation of a fovea. First, developmental studies indicate that in macaques and marmosets the foveal avascular area is defined before the fovea begins to form [18,20]. Second, optical coherence tomography in humans indicates that absence of an avascular area is associated with foveal hypoplasia, or absence of a foveal depression [14,15]. Third, finite element analysis predicts that differential elasticity in the foveal region, resulting from a local absence of retinal vessels, is sufficient to promote development of the foveal depression due to growth-induced stretch [73]. Our present results suggest that both anti-angiogenic and axon guidance factors have key roles in regulating vascular growth in the macula and in specifying the foveal region as a vascular “no go” area, and are thus indirectly implicated in the later formation of the fovea centralis.

Expression of anti-angiogenic factors in the macula during development, and guidance factors that have roles in regulating vascular growth in the macula through repellent mechanisms, suggests there may be a biological conflict of interest in the primate macula; between high density neural networks, with correspondingly high levels of energy demands, on the one hand, and adaptive pressure to constrain blood vessel growth to avoid shadowing of photoreceptors, on the other. These findings may be useful to understand the vulnerability of the macula to degeneration and to develop new therapeutic strategies to inhibit neovascularization.

Appendix 1. Levels of expression of angiogenic and anti-angiogenic genes in developing human retina


We thank the human tissue donors and the donor coordinators in the Department of Endocrinology at the Prince of Wales Hospital, Sydney, NSW, Australia, and Professor Anita Hendrickson for providing the fetal macaque tissue sections. This study was supported by grants from the Ophthalmic Research Institute of Australia to M.C.M. and J.M.P., and the ARC Centre of Excellence in Vision Science (ACEVS) Canberra, to J.M.P.; P.K. is supported by an ANU Graduate School scholarship and ACEVS; R.N. and K.B.O'B. are supported by ACEVS.


  1. Provis JM, Penfold PL, Cornish EE, Sandercoe TM, Madigan MC. Anatomy and development of the macula: specialisation and the vulnerability to macular degeneration. Clin Exp Optom. 2005; 88:269-81. [PMID: 16255686]
  2. Penfold PL, Madigan MC, Gillies MC, Provis JM. Immunological and aetiological aspects of macular degeneration. Prog Retin Eye Res. 2001; 20:385-414. [PMID: 11286898]
  3. Hageman GS, Anderson DH, Johnson LV, Hancox LS, Taiber AJ, Hardisty LI, Hageman JL, Stockman HA, Borchardt JD, Gehrs KM, Smith RJ, Silvestri G, Russell SR, Klaver CC, Barbazetto I, Chang S, Yannuzzi LA, Barile GR, Merriam JC, Smith RT, Olsh AK, Bergeron J, Zernant J, Merriam JE, Gold B, Dean M, Allikmets R. A common haplotype in the complement regulatory gene factor H (HF1/CFH) predisposes individuals to age-related macular degeneration. Proc Natl Acad Sci USA. 2005; 102:7227-32. [PMID: 15870199]
  4. Haines JL, Hauser MA, Schmidt S, Scott WK, Olson LM, Gallins P, Spencer KL, Kwan SY, Noureddine M, Gilbert JR, Schnetz-Boutaud N, Agarwal A, Postel EA, Pericak-Vance MA. Complement factor H variant increases the risk of age-related macular degeneration. Science. 2005; 308:419-21. [PMID: 15761120]
  5. Klein RJ, Zeiss C, Chew EY, Tsai JY, Sackler RS, Haynes C, Henning AK, SanGiovanni JP, Mane SM, Mayne ST, Bracken MB, Ferris FL, Ott J, Barnstable C, Hoh J. Complement factor H polymorphism in age-related macular degeneration. Science. 2005; 308:385-9. [PMID: 15761122]
  6. Yang Z, Stratton C, Francis PJ, Kleinman ME, Tan PL, Gibbs D, Tong Z, Chen H, Constantine R, Yang X, Chen Y, Zeng J, Davey L, Ma X, Hau VS, Wang C, Harmon J, Buehler J, Pearson E, Patel S, Kaminoh Y, Watkins S, Luo L, Zabriskie NA, Bernstein PS, Cho W, Schwager A, Hinton DR, Klein ML, Hamon SC, Simmons E, Yu B, Campochiaro B, Sunness JS, Campochiaro P, Jorde L, Parmigiani G, Zack DJ, Katsanis N, Ambati J, Zhang K. Toll-like receptor 3 and geographic atrophy in age-related macular degeneration. N Engl J Med. 2008; 359:1456-63. [PMID: 18753640]
  7. Bernstein SL, Borst DE, Wong PW. Isolation of differentially expressed human fovea genes: candidates for macular disease. Mol Vis. 1995; 1:4 [PMID: 9238082]
  8. Bernstein SL, Borst DE, Neuder ME, Wong P. Characterization of a human fovea cDNA library and regional differential gene expression in the human retina. Genomics. 1996; 32:301-8. [PMID: 8838792]
  9. Sharon D, Blackshaw S, Cepko CL, Dryja TP. Profile of the genes expressed in the human peripheral retina, macula, and retinal pigment epithelium determined through serial analysis of gene expression (SAGE). Proc Natl Acad Sci USA. 2002; 99:315-20. [PMID: 11756676]
  10. Bowes Rickman C, Ebright JN, Zavodni ZJ, Yu L, Wang T, Daiger SP, Wistow G, Boon K, Hauser MA. Defining the human macula transcriptome and candidate retinal disease genes using EyeSAGE. Invest Ophthalmol Vis Sci. 2006; 47:2305-16. [PMID: 16723438]
  11. Franco EC, Finlay BL, Silveira LC, Yamada ES, Crowley JC. Conservation of absolute foveal area in New World monkeys. A constraint on eye size and conformation. Brain Behav Evol. 2000; 56:276-86. [PMID: 11251320]
  12. Engerman RL. Development of the macular circulation. Invest Ophthalmol. 1976; 15:835-40. [PMID: 824224]
  13. Provis JM, Hendrickson AE. The foveal avascular region of developing human retina. Arch Ophthalmol. 2008; 126:507-11. [PMID: 18413520]
  14. McGuire DE, Weinreb RN, Goldbaum MH. Foveal hypoplasia demonstrated in vivo with optical coherence tomography. Am J Ophthalmol. 2003; 135:112-4. [PMID: 12504716]
  15. Marmor MF, Choi SS, Zawadzki RJ, Werner JS. Visual insignificance of the foveal pit: reassessment of foveal hypoplasia as fovea plana. Arch Ophthalmol. 2008; 126:907-13. [PMID: 18625935]
  16. Sandercoe TM, Geller SF, Hendrickson AE, Stone J, Provis JM. VEGF expression by ganglion cells in central retina before formation of the foveal depression in monkey retina: evidence of developmental hypoxia. J Comp Neurol. 2003; 462:42-54. [PMID: 12761823]
  17. Stone J, Sandercoe TM, Provis JM. Mechanisms of the formation and stability of retinal blood vessels. In: Tombran-Tink J, Barnstable C, editors. Ocular Angiogenesis: Diseases, Mechanisms and Therapeutics. Totowa, NJ: Humana Press; 2005. p. 101–26.
  18. Provis JM, Sandercoe T, Hendrickson AE. Astrocytes and blood vessels define the foveal rim during primate retinal development. Invest Ophthalmol Vis Sci. 2000; 41:2827-36. [PMID: 10967034]
  19. Provis JM. Development of the primate retinal vasculature. Prog Retin Eye Res. 2001; 20:799-821. [PMID: 11587918]
  20. Hendrickson A, Troilo D, Possin D, Springer A. Development of the neural retina and its vasculature in the marmoset Callithrix jacchus. J Comp Neurol. 2006; 497:270-86. [PMID: 16705674]
  21. Ohno-Matsui K, Morita I, Tombran-Tink J, Mrazek D, Onodera M, Uetama T, Hayano M, Murota SI, Mochizuki M. Novel mechanism for age-related macular degeneration: an equilibrium shift between the angiogenesis factors VEGF and PEDF. J Cell Physiol. 2001; 189:323-33. [PMID: 11748590]
  22. Park KW, Crouse D, Lee M, Karnik SK, Sorensen LK, Murphy KJ, Kuo CJ, Li DY. The axonal attractant Netrin-1 is an angiogenic factor. Proc Natl Acad Sci USA. 2004; 101:16210-5. [PMID: 15520390]
  23. Schroeder A, Mueller O, Stocker S, Salowsky R, Leiber M, Gassmann M, Lightfoot S, Menzel W, Granzow M, Ragg T. The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC Mol Biol. 2006; 7:3 [PMID: 16448564]
  24. O'Brien KM, Schulte D, Hendrickson AE. Expression of photoreceptor-associated molecules during human fetal eye development. Mol Vis. 2003; 9:401-9. [PMID: 12949469]
  25. Wu Z, Irizarry RA. A statistical framework for the analysis of microarray probe-level data. John Hopkins University, Dept. of Biostatistics Working Papers, Working Paper 73. The Berkeley Electronic Press; 2005.
  26. Bolstad BM, Irizarry RA, Astrand M, Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics. 2003; 19:185-93. [PMID: 12538238]
  27. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics. 2003; 4:249-64. [PMID: 12925520]
  28. The Gene Ontology Consortium. Gene Ontology: tool for the unification of biology. Nat Genet. 2000; 25:25-9. [PMID: 10802651]
  29. Dennis G, , Jr Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA. DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol. 2003; 4:P3 [PMID: 12734009]
  30. Provis JM, van Driel D, Billson FA, Russell P. Development of the human retina: patterns of cell distribution and redistribution in the ganglion cell layer. J Comp Neurol. 1985; 233:429-51. [PMID: 3980779]
  31. Pfaffl MW. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res. 2001; 29:e45 [PMID: 11328886]
  32. Cornish EE, Madigan MC, Natoli R, Hales A, Hendrickson AE, Provis JM. Gradients of cone differentiation and FGF expression during development of the foveal depression in macaque retina. Vis Neurosci. 2005; 22:447-59. [PMID: 16212702]
  33. Rollin R, Mediero A, Roldan-Pallares M, Fernandez-Cruz A, Fernandez-Durango R. Natriuretic peptide system in the human retina. Mol Vis. 2004; 10:15-22. [PMID: 14737067]
  34. Tombran-Tink J, Shivaram SM, Chader GJ, Johnson LV, Bok D. Expression, secretion, and age-related downregulation of pigment epithelium-derived factor, a serpin with neurotrophic activity. J Neurosci. 1995; 15:4992-5003. [PMID: 7623128]
  35. Provis J, Kozulin P. Gene expression in central vs peripheral primate retina - defining the foveal avascular zone. J Vis. 2007; 7:31a
  36. Karakousis PC, John SK, Behling KC, Surace EM, Smith JE, Hendrickson A, Tang WX, Bennett J, Milam AH. Localization of pigment epithelium derived factor (PEDF) in developing and adult human ocular tissues. Mol Vis. 2001; 7:154-63. [PMID: 11438800]
  37. van Soest SS, de Wit GM, Essing AH, ten Brink JB, Kamphuis W, de Jong PT, Bergen AA. Comparison of human retinal pigment epithelium gene expression in macula and periphery highlights potential topographic differences in Bruch's membrane. Mol Vis. 2007; 13:1608-17. [PMID: 17893662]
  38. Li HM, Wang F, Qiu W, Liu Y, Huang Q. Expression patterns of the retinal development-related genes in the fetal and adult retina. Chin Med J (Engl). 2007; 120:1716-9. [PMID: 17935677]
  39. Tombran-Tink J, Barnstable CJ. PEDF: a multifaceted neurotrophic factor. Nat Rev Neurosci. 2003; 4:628-36. [PMID: 12894238]
  40. Dawson DW, Volpert OV, Gillis P, Crawford SE, Xu H, Benedict W, Bouck NP. Pigment epithelium-derived factor: a potent inhibitor of angiogenesis. Science. 1999; 285:245-8. [PMID: 10398599]
  41. Huang Q, Wang S, Sorenson CM, Sheibani N. PEDF-deficient mice exhibit an enhanced rate of retinal vascular expansion and are more sensitive to hyperoxia-mediated vessel obliteration. Exp Eye Res. 2008; 87:226-41. [PMID: 18602915]
  42. Volpert OV, Zaichuk T, Zhou W, Reiher F, Ferguson TA, Stuart PM, Amin M, Bouck NP. Inducer-stimulated Fas targets activated endothelium for destruction by anti-angiogenic thrombospondin-1 and pigment epithelium-derived factor. Nat Med. 2002; 8:349-57. [PMID: 11927940]
  43. Cai J, Jiang WG, Grant MB, Boulton M. Pigment epithelium-derived factor inhibits angiogenesis via regulated intracellular proteolysis of vascular endothelial growth factor receptor 1. J Biol Chem. 2006; 281:3604-13. [PMID: 16339148]
  44. Levin ER, Gardner DG, Samson WK. Natriuretic peptides. N Engl J Med. 1998; 339:321-8. [PMID: 9682046]
  45. Cao LH, Yu YC, Zhao JW, Yang XL. Expression of natriuretic peptides in rat Muller cells. Neurosci Lett. 2004; 365:176-9. [PMID: 15246543]
  46. Jin Y, Zhong YM, Yang XL. Natriuretic peptides are localized to rat retinal amacrine cells. Neurosci Lett. 2007; 421:106-9. [PMID: 17566658]
  47. Chinkers M, Garbers DL, Chang MS, Lowe DG, Chin HM, Goeddel DV, Schulz S. A membrane form of guanylate cyclase is an atrial natriuretic peptide receptor. Nature. 1989; 338:78-83. [PMID: 2563900]
  48. Pedram A, Razandi M, Hu RM, Levin ER. Vasoactive peptides modulate vascular endothelial cell growth factor production and endothelial cell proliferation and invasion. J Biol Chem. 1997; 272:17097-103. [PMID: 9202027]
  49. Pedram A, Razandi M, Levin ER. Natriuretic peptides suppress vascular endothelial cell growth factor signaling to angiogenesis. Endocrinology. 2001; 142:1578-86. [PMID: 11250939]
  50. Maragoudakis ME, Missirlis E, Karakiulakis GD, Sarmonica M, Bastakis M, Tsopanoglou N. Basement membrane biosynthesis as a target for developing inhibitors of angiogenesis with anti-tumor properties. Kidney Int. 1993; 43:147-50. [PMID: 7679456]
  51. Mundel TM, Kalluri R. Type IV collagen-derived angiogenesis inhibitors. Microvasc Res. 2007; 74:85-9. [PMID: 17602710]
  52. Suchting S, Bicknell R, Eichmann A. Neuronal clues to vascular guidance. Exp Cell Res. 2006; 312:668-75. [PMID: 16330027]
  53. La Vail MM, Rapaport DH, Rakic P. Cytogenesis in the monkey retina. J Comp Neurol. 1991; 309:86-114. [PMID: 1894769]
  54. Rapaport DH, Stone J. The topography of cytogenesis in the developing retina of the cat. J Neurosci. 1983; 3:1824-34. [PMID: 6886747]
  55. Mann IC. The development of the human eye. New York: Grune and Stratton; 1964.
  56. Hendrickson A, Provis JM. Comparison of development of the primate fovea centralis with peripheral retina. In: Sernagor E, Eglen S, Harris B, Wong R, editors. Retinal Development. Cambridge: Cambridge University Press; 2006. p. 126–49.
  57. Provis JM, van Driel D, Billson FAB, Russell P. Human fetal optic nerve: overproduction and elimination of retinal axons during development. J Comp Neurol. 1985; 238:92-100. [PMID: 4044906]
  58. Lambot MA, Depasse F, Noel JC, Vanderhaeghen P. Mapping labels in the human developing visual system and the evolution of binocular vision. J Neurosci. 2005; 25:7232-7. [PMID: 16079405]
  59. Birgbauer E, Cowan CA, Sretavan DW, Henkemeyer M. Kinase independent function of EphB receptors in retinal axon pathfinding to the optic disc from dorsal but not ventral retina. Development. 2000; 127:1231-41. [PMID: 10683176]
  60. Oster SF, Deiner M, Birgbauer E, Sretavan DW. Ganglion cell axon pathfinding in the retina and optic nerve. Semin Cell Dev Biol. 2004; 15:125-36. [PMID: 15036215]
  61. Brown A, Yates PA, Burrola P, Ortuño D, Vaidya A, Jessell TM, Pfaff SL, O'Leary DD, Lemke G. Topographic mapping from the retina to the midbrain is controlled by relative but not absolute levels of EphA receptor signaling. Cell. 2000; 102:77-88. [PMID: 10929715]
  62. Knöll B, Drescher U. Ephrin-As as receptors in topographic projections. Trends Neurosci. 2002; 25:145-9. [PMID: 11852146]
  63. McLaughlin T, Hindges R, O'Leary DD. Regulation of axial patterning of the retina and its topographic mapping in the brain. Curr Opin Neurobiol. 2003; 13:57-69. [PMID: 12593983]
  64. Huberman AD, Murray KD, Warland DK, Feldheim DA, Chapman B. Ephrin-As mediate targeting of eye-specific projections to the lateral geniculate nucleus. Nat Neurosci. 2005; 8:1013-21. [PMID: 16025110]
  65. Pfeiffenberger C, Cutforth T, Woods G, Yamada J, Rentería RC, Copenhagen DR, Flanagan JG, Feldheim DA. Ephrin-As and neural activity are required for eye-specific patterning during retinogeniculate mapping. Nat Neurosci. 2005; 8:1022-7. [PMID: 16025107]
  66. Serini G, Valdembri D, Zanivan S, Morterra G, Burkhardt C, Caccavari F, Zammataro L, Primo L, Tamagnone L, Logan M, Tessier-Lavigne M, Taniguchi M, Püschel AW, Bussolino F. Class 3 semaphorins control vascular morphogenesis by inhibiting integrin function. Nature. 2003; 424:391-7. [PMID: 12879061]
  67. Pfaff D, Fiedler U, Augustin HG. Emerging roles of the Angiopoietin-Tie and the ephrin-Eph systems as regulators of cell trafficking. J Leukoc Biol. 2006; 80:719-26. [PMID: 16864601]
  68. Lu X, Le Noble F, Yuan L, Jiang Q, De Lafarge B, Sugiyama D, Bréant C, Claes F, De Smet F, Thomas JL, Autiero M, Carmeliet P, Tessier-Lavigne M, Eichmann A. The netrin receptor UNC5B mediates guidance events controlling morphogenesis of the vascular system. Nature. 2004; 432:179-86. [PMID: 15510105]
  69. Cowan CA, Henkemeyer M. Ephrins in reverse, park and drive. Trends Cell Biol. 2002; 12:339-46. [PMID: 12185851]
  70. Curcio CA, Sloan KR, Kalina RE, Hendrickson AE. Human photoreceptor topography. J Comp Neurol. 1990; 292:497-523. [PMID: 2324310]
  71. Putnam NM, Hofer HJ, Doble N, Chen L, Carroll J, Williams DR. The locus of fixation and the foveal cone mosaic. J Vis. 2005; 5:632-9. [PMID: 16231998]
  72. Adams DL, Horton JC. Shadows cast by retinal blood vessels mapped in primary visual cortex. Science. 2002; 298:572-6. [PMID: 12386328]
  73. Springer AD, Hendrickson AE. Development of the primate area of high acuity, 3: temporal relationships between pit formation, retinal elongation and cone packing. Vis Neurosci. 2005; 22:171-85. [PMID: 15935110]