Differential distribution of fibroblast growth factor ... · The pedicles of developing and adult cones are FGFR1 and FGFR3 immunoreactive, and the basal, synaptic region is FGFR4

Molecular Vision 2004; 10:1-14 <http://www.molvis.org/molvis/v10/a1>Received 28 May 2003 | Accepted 7 November 2003 | Published 8 January 2004

Cone photoreceptors are dominant in the fovea centralisof primates where they reach a peak density of up to 300 K/mm2 [1]. In the fovea and parafovea, cones are the source ofneural pathways mediating both high acuity and colour vision[2-4]. During development, the site of the future foveal de-pression (the “incipient fovea”) is the first region to differen-tiate, and cone photoreceptors are amongst the earliest cellsgenerated (from around fetal day (Fd) 38 in macaques) [5].Initially, cones are identifiable as cuboidal, epithelial-like cellsthat subsequently differentiate morphologically over an ex-tensive period. Foveal cones are involved in synaptic struc-tures shortly after differentiation [6,7] (reviewed in [8]) withsmall developing outer segments evident from fetal day (Fd)75 in macaques [9] and about 11 weeks gestation in humans.However, foveal cones achieve final, adult-like proportions at

about one year postnatal in macaques and several years post-natal in humans [10-12].

Two specialized features of adult foveal cones are (1) theirgreatly elongated axons (the fibers of Henle), and (2) theirnarrow, elongated inner and outer segments, when comparedwith cones outside the fovea. The fibers of Henle enable syn-aptic contact between cones in the fovea with bipolar and hori-zontal cells that are displaced onto the foveal rim during for-mation of the foveal depression, which takes place in macaqueretina between Fd 105 and around birth, at Fd 172 [10,11,13-16]. In adult retina, the narrow and elongated morphology offoveal cones is correlated with high cone density, the mostnarrow and elongated cones occurring at the site of peak den-sity, in the foveola [1,15,17]. During development, progres-sive increase in foveal cone density is correlated with a de-crease in cone diameter [14,15,18,19]. These observations leadto the suggestion that narrowing and elongation of foveal conesis the mechanism by which cone photoreceptors crowd into,or are displaced towards, the fovea. However, the mechanismseffecting the narrowing and elongation of cones have not beenexplored.

©2004 Molecular Vision

Differential distribution of fibroblast growth factor receptors(FGFRs) on foveal cones: FGFR-4 is an early marker of conephotoreceptors

Elisa E. Cornish,1 Riccardo C. Natoli,1 Anita Hendrickson,2 Jan M. Provis1,3

1Save Sight Institute and 3Department of Anatomy & Histology, University of Sydney, New South Wales, Australia; 2Department ofBiological Structure, University of Washington, Seattle, WA

Purpose: Relatively little is known of the expression and distribution of FGF receptors (FGFR) in the primate retina. Weinvestigated expression of FGFRs in developing and adult Macaca monkey retina, paying particular attention to the conerich, macular region.Methods: One fetal human retina was used for diagnostic PCR using primers designed for FGFR1, FGFR2, FGFR3,FGFR4, and FGFR like-protein 1 (FGFrl1) and for probe design to FGFR3, FGFR4, and FGFrl1. Rat cDNA was used tosynthesize probes for FGFR1 and FGFR2 with 90% and 93% homology to human, respectively. Paraffin sections of retinafrom macaque fetuses sacrificed at fetal days (Fd) 64, 73, 85, 105, 115, 120, and 165, and postnatal ages 2.5 and 11 yearswere used to detect FGF receptors by immunohistochemistry and in situ hybridization.Results: PCR showed each of the FGF receptors are expressed in fetal human retina. In situ hybridization indicated thatmRNA for each receptor is expressed in all retinal cell layers during development, but most intensely in the ganglion celllayer (GCL). FGFR2 mRNA is reduced in the adult inner (INL) and outer (ONL) nuclear layers, while FGFrl1 mRNA isvirtually absent from the adult ONL. FGFR4 mRNA is particularly intense in fetal and adult cone photoreceptors. Immu-noreactivity to FGFR1-FGFR4 was detected in the interphotoreceptor matrix in what appeared to be RPE microvilliassociated with developing photoreceptor outer segments, and generally is high in the GCL and low in the INL. Differentpatterns of FGFR3 and FGFR4 immunoreactivities in the outer plexiform layer (OPL) suggest localization of FGFR3 tohorizontal cell processes, with FGFR4 being expressed by both horizontal and bipolar cell processes. FGFR1, FGFR3,and FGFR4 immunoreactivities are present in the inner segments and somata of adult cones. The pedicles of developingand adult cones are FGFR1 and FGFR3 immunoreactive, and the basal, synaptic region is FGFR4 immunoreactive.FGFR4 labels cones almost in their entirety from early in development and is not detected in rods. The fibers of Henle areintensely FGFR4 immunoreactive in adult cones.Conclusions: The results show high levels of FGF receptor expression in developing and adult retina. Differential distri-bution of FGF receptors across developing and adult photoreceptors suggests specific roles for FGF signalling in develop-ment and maintenance of photoreceptors, particularly the specialized cones of the fovea.

Correspondence to: Jan M. Provis, Department of Anatomy & His-tology (F13), University of Sydney, NSW 2006, Australia; Phone:+61 2 9351 4195; FAX: +61 2 9351 2813; email:[email protected]

1

The mammalian fibroblast growth factor (FGF) familycomprises at least 23 structurally related polypeptides that in-teract with low affinity heparan sulfate proteoglycans mol-ecules to activate high affinity, transmembrane, FGF tyrosinekinase receptors (FGFR) [20,21]. There are at least four suchreceptors [22-24], with two additional receptors, FGFR5 orFGF receptor-like protein-1 (rl-1) [25-27] and FGFR6 [28]also proposed. FGFs have a diverse range of functions in thecentral nervous system (reviewed in [29]), including roles inproliferation [30-33], differentiation and survival [34-37], andaxon guidance [38,39].

While a considerable number of studies have investigatedthe effects of FGFRs in cell lines, transgenic models and knock-outs [40-43], relatively little is understood about the normaldistribution and significance of FGFRs in the mammalian ner-vous system, including the retina. FGFR distributions havebeen described in chicken [44-47], bovine [48,49], and rat andhuman retinas [50-53]. However, few studies have made athorough analysis of the specific cell classes expressing dif-ferent FGFRs and relatively little attention had been paid tothe FGFRs expressed by cone photoreceptors, especially dur-ing development (see however [52-54]). In view of the sig-nificant neuroprotective effects of FGFs in the retina [55-58]a systematic analysis of expression of FGFRs is of interest.Furthermore, in view of the known morphogenetic effects ofFGF2 in particular [51,59-63] we aimed first, to determine ifFGFRs are expressed in the developing human and monkeyretina using PCR, and second, to investigate by in situ hybrid-ization and immunohistochemistry if they are expressed bycone photoreceptors in patterns that may be consistent with arole for FGFs in foveal cone morphogenesis.

METHODSDiagnostic PCR and primer design: Primer pairs were de-signed for the four Fibroblast Growth Factor high affinity re-ceptors (FGFR1, FGFR2, FGFR3, FGFR4) and the Fibroblas-tic Growth Factor Receptor Like-1 (FGFrl1, also know asFGFR5). Due to the presence of highly conserved regionsbetween receptors, particular care was taken to ensure thatonly one receptor was amplified by each pair of primers (Table1). PCR products were analysed using both restriction diges-tion and sequencing.

Owing to difficulties obtaining fresh macaque retinas, weused human fetal retina for RT-PCR and probe design forFGFR3, FGFR4, and FGFrl1. One human fetal eye at 19 weeks

gestation was obtained with informed maternal consent andapproval of the Human Ethics Committee of the University ofSydney. The retina was dissected free of the choroid and reti-nal pigmented epithelium and RNA extracted using TriZOL®(GIBCO-BRL, Invitrogen, Sydney, Australia) for expressionanalysis. Single stranded cDNA was created from the extractedretinal RNA using Reverse Transcriptase-Polymerase ChainReaction (RT-PCR) as per Reverse Transciption System Kit(Promega, Sydney, Australia).

Using the human fetal cDNA as a template, target se-quences (as given in Table 1) were PCR amplified using thefollowing solutions (final concentrations in brackets) and pro-tocol parameters. PCR Master solution mix (Promega) con-taining 10.8 µl of RNase free H

20, 2 µl of 10X buffer, 2 µl of

10 mM dNTPs (1 mM), 2 µl of 25 mM MgCl2 (1 mM), 0.2 µl

of 5 U/µl Taq DNA Polymerase in storage buffer B (Promega),1 µl of 40 ng/µl cDNA (40 ng), 1 µl of 1.25 mM primer 1(1.25 mM; Sigma-Genosys, Sydney, Australia), 1 µl of 1.25mM primer 2 (1.25 mM; Sigma-Genosys) to make a final vol-ume of 20 µl per reaction mix. PCR protocol: DNA was dena-tured at 94 °C for 3 min, then amplified over 40 cycles at 94°C for 25 s, 58 °C for 25 s, 72 °C for 25 s, followed by a finalextension step at 72 °C for 5 min. PCR products were run ona 3% agarose gel containing 0.2 ng per ml of ethidium bro-mide (Bio-Rad, Sydney, Australia) with two standard lanes(Hyperladder-Bioline, Astral, Sydney, Australia). The gel wasviewed on a UV transilluminator, photographed on Polaroidfilm and scanned to produce a digital image.

RNA probes: Complementary DNA from rat used in an-other study (donated by Andrew Baird, Scripps Institute) wasused to prepare RNA probes for FGFR1 (flg) and FGFR2 (bek).Sequencing showed these probes to be 90% and 93% homolo-gous with human FGFR1 and FGFR2, respectively (Genbankaccession numbers as per Table 1). We have no data on thesequences for the FGF receptors in macaque, but sequencealignments for each receptor across various species showFGFR to be highly conserved [24,64]. Because fresh macaquetissue was not available to us, we used sequenced DNA frag-ments obtained from fetal human retina to generate RNAprobes for FGFR3, FGFR4, and FGFrl1. DNA fragments wereamplified from total RNA of human fetal retina by PCR, li-gated to pGem®-T DNA vector (Promega catalog numberA3610) and cloned in JM109 competent cells, using thePromega pGem® T DNA vector system protocol. Antisenseand sense probes for FGFR1, FGFR2, FGFR3, FGFR4, and

©2004 Molecular VisionMolecular Vision 2004; 10:1-14 <http://www.molvis.org/molvis/v10/a1>

TABLE 1. PRIMERS USED FOR FGF RECEPTOR GENE PCR

Primers (5' to 3') ---------------------------------------------- Product Gene Forward primer Reverse primer Accession number size (bp)------ -------------------- ----------------------- -------------------------------------- ---------FGFR1 taccaccgacaaagagatgg ctggctgtggaagtcactct NM_000604: 2825 - 2845 and 3089 - 3109 287FGFR2 tggagcgatcgcctcaccg cttccaggcgctggcagaactgt NM_000141: 2675 - 2693 and 3001 - 3023 352FGFR3 caccaccgacaaggagcta gctcgagctcggagacatt XM_044120: 1398 - 1416 and 1810 - 1828 433FGFR4 gggtcctgctgagtgtgc ggggtaactgtgcctattcg XM_030308: 556 - 573 and 940 - 959 406FGFrl1 cggctcctacctcaataagc aacgagggaaggtccttgt NM_021923: 970 - 989 and 1308 - 1326 359

The primer pairs used to identify FGF receptors in human retina by PCR, and their accession numbers, are shown below.

2


Figure 1. Distribution of FGFR1, FGFR2, and FGFR3 mRNAs in fetal and postnatal monkey retina. Developmental series of macaque retinalsections hybridized using rat RNA probes to show expression of FGFR1 (A-D) and FGFR2 (E-H) mRNA (red); and FGFR4 mRNA expres-sion (I-K) using a probe designed from fetal human PCR product. Some sections are counter-labelled (green) to show either L/M opsin (A, E)or vimentin-IR. A: The incipient fovea, showing cytoplasmic FGFR1 mRNA expression in all cellular layers and in L/M opsin-IR cones.Horizontal arrowhead (A, E) indicates the position of the developing outer plexiform layer. B: Intense FGFR1 mRNA expression in thecytoplasm of foveal ganglion cells (red) and the end feet of MC, which are also vimentin-IR (yellow, arrowheads). FGFR1 mRNA is alsodetected on the outer aspect of the cone photoreceptor nuclei and in the adjoining inner segment, separated by the external limiting membrane(horizontal arrow). C: FGFR1 mRNA is low in the INL in postnatal retina. D: FGFR1 sense probe hybridized to adult retina, counter labelledwith anti-vimentin (green). E: FGFR2 in all cell layers at the incipient fovea including L/M opsin-IR cones. F: Intense FGFR2 expression inthe GCL, inner segments of cones and inner and outer aspects of the INL, during formation of the foveal depression. G: FGFR2 expression ishigh in the GCL in postnatal retina but low in the INL and ONL. H: FGFR2 sense probe. I: Co-localization of FGFR4 with vimentin-IR Müllercells (yellow, arrowheads). Cones express abundant FGFR4 mRNA. J: FGFR4 expression is strongest in the GCL and ONL in adult retina. K:FGFR4 sense probe, immunolabelled with anti-vimentin. All scale bars represent 20 µm. The ganglion cell layer (GCL), inner nuclear layer(INL), inner plexiform layer (IPL), and outer nuclear layer (ONL) are also identified.

3

FGFrl1 were generated by in vitro transcription of linearizedplasmid constructs containing the FGFRs using bacterioph-age RNA polymerases (Promega). The templates were puri-fied, sequenced then RNA probes prepared using Digoxigenin(DIG)-labelled-11-UTP using DIG RNA labelling kit (Roche,Sydney, Australia).

Monkey tissue: Macaque monkey retinae aged Fd 64,73, 85 (before development of the fovea), 105, 115, 120, and164 (during development of the fovea) and postnatal (P) 2.5and 11 years (after development of the fovea) were obtained

with the ethical approval of the University of Washington,Seattle, from Bogor Agricultural University, Indonesia. Allanimal procedures were in accordance with guidelines estab-lished by the NIH. Fetuses were delivered by caesarean sec-tion and mothers returned to the breeding colony after recov-ery. Fetuses were euthenased by an intravascular overdose ofbarbiturate, eyes enucleated, then injected with methylCarnoy’s fixative and returned to fixative for 2-4 h. Eyes wereembedded in paraffin and sectioned in the horizontal plane at8 µm.


Figure 2. Distribution of FGFR3 and FGFrl1 mRNAs in fetal and postnatal monkey retina. Sections of fetal and adult monkey retina hybrid-ized using human RNA probes to show expression of FGFR3 (A-C) or FGFrl1 (D-F) mRNA (red) and vimentin-IR (green). A: Cytoplasmicand nuclear FGFR3 mRNA is abundant in virtually all cells in the developing retina. Expression in Müller cell somata could not be ruled out,but FGFR3 mRNA was not detected in Müller cell processes. Horizontal arrowhead indicates the position of the developing OPL. B: FGFR3mRNA is predominantly in ganglion cells in the adult. C: FGFR3 sense probe. The apparent labelling of the RPE is autofluorescence; somecapillaries bound the sense probe but neural elements were not labelled. D: FGFrl1 mRNA is expressed in ganglion cells in the developingfovea, in some INL cells, including Müller cells that co-localize vimentin (oblique arrowhead), and the inner segments of photoreceptors,including cones. E: FGFrl1 expression is virtually confined to the GCL in the adult. F: FGFrl1 sense probe. Apparent labelling in the RPE isautofluorescence. All scale bars represent 20 µm. The ganglion cell layer (GCL), inner nuclear layer (INL), inner plexiform layer (IPL), outernuclear layer (ONL), outer plexiform layer (OPL), and external limiting membrane (elm) are also identified.

4

In situ hybridization: Sections were de-waxed in twochanges of xylene for 10 min each, then hydrated in gradedethanols and rinsed in PBS (0.9% NaCl in phosphate buffer,pH 7.4). They were then fixed in 10% Neutral Buffered For-malin (NBF) for 20 min, washed in PBS for 5 min, and thenplaced in a solution containing 20 mg/ml of proteinase K di-luted in TE (50 mM Tris-HCl, 5 mM EDTA, pH 8) for 7 minat 37 °C. The sections were then rinsed in PBS, and re-fixedin NBF for 20 min. Slides were placed into a solution contain-ing 0.1 M triethanolamine (pH 8.0) and acetic anhydride(Sigma catalog number A-6404) for 10 min, washed in PBSthen 0.9% NaCl for 5 min each. The slides were dehydratedusing graded ethanols and air dried.

The pre-hybridization solution was pre-heated to over 65°C, before being added to the section, and incubated at 60 °Cfor a minimum of 1 h under a coverslip. Following incuba-tion, the coverslip and pre-hybridization solution were care-fully removed and the pre-heated hybridization solution con-taining the probe at its particular concentration was addedunder a new coverslip. Sections were hybridized overnight, attemperatures optimized for each probe. Coverslips were re-moved and sections washed for 5 min at room temperature

(RT) in 2X saline sodium citrate (SSC, pH 7.4), 0.5X SSCand 0.1X SSC, followed by 0.1X SSC at the appropriate post-hybridization wash temperature (55 °C to 75 °C) for 2 h. Post-hybridization temperature was determined as that which pro-duced the lowest levels of background anti-sense labelling,and the cleanest sense labelling. The slides were then washedin 0.1X SSC at RT for 5 min.

Slides were rinsed in washing buffer, placed into block-ing solution for 30 min, then incubated in 1:2000 anti-DIGantibody for 1 h. Following antibody incubation, slides wererinsed in two changes of washing buffer then rinsed in detec-tion buffer for 5 min. Visualisation of DIG-labelled probeswas performed using Roche HNPP “Fast Red” fluorescent la-bel, incubated for no longer than 1 h. Color reaction wasstopped in Milli Q-water (MQ-H

20) for 30 min. Sections that

were not counter labelled, were fixed in NBF for 20 min,washed in PBS and then coverslipped with DABCO (6 g/L,Sigma; 20% PBS/glycerol) and sealed with nail varnish.

For counter immunolabelling, hybridized sections wereblocked for 30 min in normal goat serum (NGS) in prepara-tion for immunolabelling. Sections were incubated overnightat 4 °C in anti-vimentin (1:100; mouse anti-swine, DAKO) or


Figure 3. FGFR1 immunoreactivity in monkey fetal and postnatal central retina. FGFR1-IR (red) in fetal and adult macaque retina, doublelabelled with anti-CRALBP (green; A, B), anti-vimentin (green; C-G) and anti-synaptophsyin (blue; D, G). A: FGFR1-IR is present in axonalbundles in the nerve fiber layer (NFL), on developing pedicles of cone photoreceptors (asterisk) and presumed developing microvilli (dmv) ofthe retinal pigmented epithelium (RPE). B: FGFR1-IR associated with the developing cone pedicles (asterisk) and on the presumed develop-ing microvilli (dmv). C: Intense FGFR1-IR is present at the level of the developing cone pedicles (asterisk), in the inner segments of cones andassociated with in the microvilli (dmv) of the RPE. D: High magnification of the ONL at Fd 120 showing synaptophysin-IR in cones. E:Foveal rim. FGFR1-IR is intense in the GCL and IPL, on cone pedicles (asterisk), inner segments, RPE and to a lesser degree, cone somata.The green labelling in the RPE and at the level of the IS and OS is non-specific. F: A high power view showing FGFR1-IR cone pedicles. G:The same section as F, showing co-localization of FGFR1 protein with synaptophysin on the base of the pedicle. All scale bars represent 20µm. The cone pedicle (CP), ganglion cell layer (GCL), inner nuclear layer (INL), inner plexiform layer (IPL), inner segment (IS), presumeddeveloping microvilli (dmv), nerve fiber layer (NFL), retinal pigmented epithelium (RPE), ellipsoid (e), external limiting membrane (elm),myoid (m), and nuclei (n) are also identified.

5

anti-CRALBP (1:1000, courtesy Jack Saari, University ofWashington, Seattle). After three 10 min washes in PBS, theantibody was detected using Alexa goat anti-mouse 488(1:1000, Molecular Probes, Invitrogen, Sydney, Australia), thenrinsed, coverslipped with DABCO (6 g/L, Sigma) in 20% PBS/glycerol and sealed with nail varnish. Negative control ex-periments were performed for each antibody on sections fromeach animal used in the study, either by omitting the primaryantibodies or by using a non-immuno isotype control.

Immunohistochemistry: Paraffin sections were de-waxedand rehydrated, as described above, then rinsed in PBS twicefor 10 min. Antigen unmasking involved heating sections in10 mM sodium citrate butter (pH 6.0) at 80 °C for 10 min thencooling to RT before washing in PBS for 10 min. The areaaround each section was dried and circled with a DAKO pento contain solutions and to ensure uniform coverage of thesections. To reduce non-specific staining, sections were ini-tially incubated for 30 min on the shaker table at RT, in 10%NGS, 0.04% saponin in PBS. Sections were then incubated inthe FGFR antibody (anti-FGFR1, anti-FGFR2, anit-FGFR3,anti-FGFR4; rabbit anti-human; 1:200 except FGFR4 at 1:400;

Santa Cruz, Biotechnology, Inc.) with 0.04% saponin and 2%NGS/PBS for 48-60 h in a humidity chamber at 4 °C, thenrinsed with 2% fetal bovine serum (FBS) in PBS for 5 min.After rinsing, sections were incubated in goat anti-rabbit IgG-conjugated Alexa 594 (Molecular Probes; 1:1000) in PBS for40 min, then washed with 2% FBS/PBS for 15 min. From thispoint, all incubations and washes were performed covered.Second primary antibodies were diluted in 0.04% saponin, 2%NGS/PBS and incubated overnight at 4 °C. Second antibod-ies were as follows: Anti-rhodopsin (1:100 Rho4D2; mono-clonal mouse courtesy Dr. Robert Molday, University of Brit-ish Columbia, Canada); anti-vimentin (1:100, mouse anti-swine, DAKO); anti-cellular retinaldehyde-binding protein(1:100 CRALBP courtesy Dr. Jack Saari, University of Wash-ington, Seattle); and α-transducin (1:100, mouse anti-human;[65]). Sections were incubated in goat anti-mouse IgG-conju-gated Alexa 488 (Molecular Probes; 1:1000) in PBS for 40min at RT to visualize bound antibody. Antibodies were raisedagainst human sequences and specificity for monkey FGFRshas not been proven conclusively.


Figure 4. FGFR2 immunoreactivity in monkey fetal and postnatal central retina. FGFR2-IR (red) in fetal and adult macaque retina, doublelabelled with anti-CRALBP (green; A, B), anti-vimentin (green; C-E) and anti-synaptophsyin (blue; E). A: Low FGFR2 immunoreactivity inall retinal layers and in the RPE at the incipient fovea. Horizontal arrowhead shows the position of the developing OPL (A,B). B: A highmagnification of the outer retina and RPE. C: FGFR2-IR is intense in the GCL on the rim of adult fovea, and at lower levels in the INL andONL. The green labelling in the RPE and at the level of the IS and OS is non-specific. D: A high power view of the OPL in adult retina showingweak FGFR2-IR in the OPL. E: There is a virtual absence of FGFR2 on the cone pedicles and axons. All scale bars represent 20 µm. The conepedicle (CP), ganglion cell layer (GCL), inner nuclear layer (INL), inner plexiform layer (IPL), inner segment (IS), presumed developingmicrovilli (dmv), nerve fiber layer (NFL), retinal pigmented epithelium (RPE), blood vessel (bv), and cone nuclei (n) are also identified.

6

Some sections were triple labelled to demonstrate synap-tic terminals, after confocal imaging of the initial double la-bel. After 3 washes in PBS for 15 min each these sectionswere incubated in anti-synaptophysin (rabbit anti-human,1:200, DAKO) in PBS overnight. Following several washesin PBS, sections were then incubated in goat biotinylated anti-rabbit secondary antibody for 45 min then washed andvisualised using streptavidin-conjugated Cy5 (1:100).

Analyses: All retinae were viewed using a scanning LeicaConfocal Microscope and TCSNT software, version 1.6.587and prepared using Adobe Photoshop version 5.5. Images werecaptured using a 16x (field size 512x512 µm) or 40x objec-tive (field size 250x250 µm) and a selection of images werezoomed at 2 or 4 times.

RESULTSFGFR PCR and in situ hybridization.: RT-PCR showed ex-pression of each of the five receptors in human fetal retina.Sense RNA probes for each of the FGFRs produced no de-

tectable labelling in control sections under optimal hybridiza-tion conditions. In contrast, distinct patterns of mRNA expres-sion were discerned using each of the FGFR antisense probes(Figure 1, Figure 2).

The ganglion cell layer (GCL) was heavily labelled usingRNA probes for each of the FGFR at all ages (Figure 1, Fig-ure 2). In most cases labelling was cytoplasmic, with the ex-ception of FGFR3 probe, which showed both cytoplasmic andnuclear localization (Figure 2A,B). In fetal retina, the innernuclear layer (INL) was labelled using each of the FGFRprobes, but labelling for FGFR2, FGFR3, and FGFrl1 in theINL declined with age (Figure 1A,B,E,F) so that little expres-sion of FGFR2 and FGFrl1 receptors was detected in the adultINL (Figure 1G, Figure 2E) and only low levels of FGFR3were present (Figure 2B). Expression of each of the FGFRswas detected in the outer nuclear layer (ONL) of fetal retinae(Figure 1, Figure 2). In adult retina, FGFR2 expression in theONL was at very low levels (Figure 1G) and FGFrl1 is virtu-ally absent (Figure 2E).


Figure 5. FGFR3 immunoreactivity in monkey fetal and postnatal central retina. FGFR3-IR (red) in fetal and adult macaque retina,immunolabelled with anti-vimentin (green; A-E) and anti-synaptophsyin (blue; E). A: FGFR3-IR on the foveal rim. Müller cell somata andprocesses colocalize vimentin and FGFR3 (oblique arrowheads). The OPL is almost completely labelled except for a narrow band adjacent tothe cone pedicles (double asteriks). B: A high magnification showing FGFR3-IR in the INL, OPL, and ONL, and indicating (double asterisks)the non-reactive band. Developing outer segments of some cones are also FGFR3-IR. C: FGFR3-IR is intense in the majority of cells in theGCL, in isolated cells in the INL, in the majority of ONL somata, in the inner segments of photoreceptors as well as in the IPL and OPL on thefoveal rim. The proximal parts of FH are also immunoreactive. The green labelling in the RPE and at the level of the IS and OS is non-specific.D: A high power view showing FGFR3-IR in the OPL and on the bases of the cone pedicles. A non-reactive band, about 2 µm deep, is indicated(double asterisks) E: The same section after labelling with showing co-localization of FGFR3 with synaptophysin on the bases of the pedicles.All scale bars represent 20 µm. The cone pedicle (CP), developing outer segments (dOS), fibers of Henle (FH), ganglion cell layer (GCL),inner nuclear layer (INL), inner plexiform layer (IPL), and nuclei (n) are also identified.

7

Double labelling with anti-L/M opsin at Fd 95 showedco-localization with FGFR1 and FGFR2 mRNA (Figure 1A,E)in cone photoreceptors; similar results were found for FGFR3,FGFR4, and FGFrl1 (not shown). From around Fd 120, pho-toreceptors showed expression of FGFR1, FGFR2, and FGFR4mRNA in the developing inner segments, along with somecytoplasmic expression (Figure 1). Postnatally, FGFR1 andFGFR4 were the most strongly expressed FGFRs in cone in-ner segments (Figure 1C,J). Double labelling using anti-vimentin, a useful marker of Müller cells, indicated that bothFGFR1 and FGFR4 are expressed by Müller cells, and thatthis expression is high during formation of the foveal depres-sion (Figure 1B,I, oblique arrowheads).

Immunoreactivity: FGFR1 immunoreactivity (-IR) wasdetected throughout the retina at Fd 85 but was most intensein axon bundles in the nerve fiber layer (NFL) and in conephotoreceptors at the level of the developing pedicles (Figure3A,B, asterisks), on the outer aspect of the nucleus, and onmembranes in the subretinal space that appear to be RPE mi-crovilli (Figure 2A,B). These membranes did not colocalize

with either anti-CRALBP or anti-vimentin. An attempt to la-bel sections with an antibody to interphotoreceptor retinoidbinding protein (IRBP; polyclonal antibody courtesy of Dr.G. Chader [66]) was unsuccessful in the methyl Carnoy-par-affin material (data not shown). However, based on the rela-tionship of these membranes to the outer aspect of the devel-oping cones, we tentatively identify them as developing RPEmicrovilli. A similar pattern of immunoreactivity in outer retinais more pronounced by Fd 120, when there is a distinct bandof FGFR1-IR at the level of the developing outer plexiformlayer (OPL; Figure 3C, asterisk) although triple labelling withanti-synaptophysin did not show co-localization with FGFR1in synaptic pedicles at this stage of development (Figure 3D,asterisk). Intense FGFR1-IR was present also in the inner seg-ments of cones and presumed RPE microvilli that invest thedeveloping outer segments (Figure 2D).

Unlike fetal retina, FGFR1-IR was strong in the GCL andinner plexiform layer (IPL) of adult retina (Figure 3E). Im-munoreactivity was also strong on the membranes of semicir-cular structures adjacent to the OPL, identified as cone pedicles


Figure 6. FGFR4 immunoreactivity in monkey fetal and postnatal central retina. FGFR4 immunoreacitivity (red) in fetal and adult macaqueretina, immunolabelled with anti-vimentin (green; A-E) and anti-synaptophsyin (blue; E). A: FGFR4-IR in ganglion cell somata, processesand in cones on the edge of the developing fovea. B: A high magnification of the outer retina at the developing fovea showing FGFR4-IR conesomata and pedicles (asterisk). C: FGFR4-IR on the rim of adult fovea is intense in the cytoplasm of ganglion cells and in presumed ganglioncell dendrites in the IPL. FGFR4-IR is present in the OPL, along the length of fibers of Henle and in the inner segments of photoreceptors. Thegreen labelling in the RPE and at the level of the IS and OS is non-specific. D: A high power view in adult retina shows FGFR4-IR in the fibersof Henle, and mild IR associated with the bases of cone pedicles and in the OPL. E: Synaptophysin-IR colocalizes with FGFR4 on the basesof the pedicles. All scale bars represent 20 µm. The cone pedicle (CP), fibers of Henle (FH), ganglion cell layer (GCL), inner nuclear layer(INL), inner plexiform layer (IPL), outer plexiform layer (OPL), presumed developing microvilli (dmv), and external limiting membrane(elm) are also identified.

8

(Figure 3E, asterisk; Figure 4F,G), although the fibers of Henlewere not FGFR1-IR (Figure 3E). Triple labelling of adult retinashows synaptophysin immunoreactivity (blue) on the base ofthe pedicle colocalized with FGFR1 (Figure 3G). The RPE,cone somata, and IS were also FGFR1-IR in adult retina (Fig-ure 2E,F).

FGFR2-IR at Fd 85 was weak in both the retina and RPE(Figure 4A,B). In adult retina, FGFR2-IR was moderate inthe INL and GCL but weak in the ONL (Figure 4C) and virtu-ally absent from cone pedicles (Figure 4C-E).

A striking feature of the pattern of FGFR3-IR was thepresence of labelling throughout both plexiform layers (Fig-ure 5). The GCL was intensely immunoreactive in both devel-oping and adult retina and it is likely, therefore, that IPL-IRincludes the dendrites of ganglion cells. Cells that appeared tobe “displaced” ganglion cells in the INL, with dendrites enter-ing the IPL, also were identified at both Fd 120 and in adultretina (Figure 5A,C). Weak IR in the presumed RPE microvilliwas seen in fetal retina (Figure 5A). There is some evidenceof co-localization of FGFR3 with vimentin in adult retina (Fig-ure 5C, oblique arrowhead) but this was more pronounced inthe developing fovea, where Müller cells were clearly doublelabelled (Figure 5A, oblique arrowheads).

Some presumed horizontal cells in the adult retina wereFGFR3-IR (not shown) and it is possible that these are thesource of the FGFR3-IR processes spreading throughout mostof the depth of the OPL. A prominent feature of the OPL la-belling in fetal retina is a distinct non-reactive band, about 2µm wide, adjacent to the cone pedicles (Figure 5A,B, doubleasterisks). A similar feature, although not quite so clearly de-fined, was detected in adult retina abutting the bases of thecone pedicles, which are also FGFR3-IR (Figure 5D). Triplelabelling showed synaptophysin immunoreactivity colocalizedwith FGFR3 on the bases of the cone pedicles, but absent fromthe non-reactive band (Figure 5E, double asterisks), suggest-

ing that the band comprises post-synaptic elements. Cone so-mata and inner segments wre FGFR3-IR in fetal and adultretina, as were the proximal parts of the fibers of Henle inadult retina (Figure 5C).

FGFR4-IR in fetal retina was intense in cone photorecep-tors and in the processes of ganglion cells, which could beidentified clearly within the GCL and IPL (Figure 6A). Co-localization of FGFR4 and vimentin in the inner retina indi-cates presence of FGFR4 on the inner processes of Müllercells, particularly in the developing fovea (Figure 6A, yellowlabelling, oblique arrowhead). Cone photoreceptors were la-belled in their entirety in fetal retina, including punctate la-belling at the level of the cone pedicles (Figure 6B, asterisk).Presumed RPE microvilli also were mildly FGFR4-IR. A simi-lar pattern of immunoreactivity was seen in adult retina, in-cluding FGFR4-IR along the full length of the fibers of Henle,but was reduced significantly on the cone pedicles (Figure6C). The OPL was moderately FGFR4-IR (Figure 6D), wheresignal colocalized with synaptophysin (Figure 6E).

Because of the potential value of FGFR4 as a specific,early marker of cones, we investigated FGFR4-IR in speci-mens at Fd 64, 73, 85, and 105, which were double labelledusing antibodies to rhodopsin (to label rods) or antibody to α-transducin, which is specific to cones [65]. Cuboidal cells inthe ONL at the incipient fovea were FGFR4-IR at Fd 64 (Fig-ure 7A), that is, about 10 days before L/M opsin can be de-tected by immunohistochemistry or in situ hybridization incentral macaque retina [9]. At Fd 73, many cells in the ONLwere FGFR4-IR; a small number of rhodopsin positive cellswere detected near the incipient fovea, but none co-localizedFGFR4-IR (Figure 7B). Foveal cones co-localizing FGFR4-IR and α-transducin-IR were detected at Fd 85 (Figure 7C).At Fd 105, double labelling with anti-rhodopsin showed manyphotoreceptors immunoreactive to either rhodopsin or FGFR4,but none that co-localized the two markers (Figure 7D).


Figure 7. Cones, but not rods, are FGFR4 immunoreactive in fetal monkey retina. FGFR4-IR (red) in sections double labelled with anti-rhodopsin (green; B, D) and antibody to the cone-specific marker, α-transducin (green; C). A: Cones in the incipient fovea are stronglyimmunoreactive to FGFR4 at Fd 64, when very few cones are immunoreactive to L/M opsin. B: A slightly oblique section on the edge of theincipient fovea shows a small number of rhodopsin-IR cells in the ONL which did not colocalize with FGFR4. C: FGFR4-IR cells colocalizedwith the cone marker, α-transducin. D: FGFR4-IR cells had a distinctive cone morphology and were interspersed with rhodopsin-IR rods onthe edge of the incipient fovea. No cells colocalized rhodopsin and FGFR4. All scale bars represent 10 µm. The cone nuclei (n) and a rod (r)are also identified.

9

DISCUSSION This study provides a detailed description of the expressionpatterns of FGF receptor mRNA and protein in developingand adult primate retina. Owing to the unavailability of freshmacaque retinal tissue, we used cDNA from rat to synthesizeRNA probes for FGFR1 (90% homolous with human) andFGFR2 (93% homologous), and PCR products from freshhuman retina to design RNA probes for FGFR3, FGFR4, andFGFrl1. We have assumed specificity of these probes for FGFreceptor mRNAs in macaque tissue, even though specificityhas not been confirmed for macaque. Considerable attentionhas been paid previously to the distribution of FGFR1 andFGFR2 in normal developing and adult retina [67-71], in reti-nal detachment [72,73], and in several other experimental para-digms [58,74-79]. However, until recently, little was knownabout the expression of FGFR3 and FGFR4 in the retina[52,53]. Those recent studies have reported expression ofFGFR4 mRNA in human retina [52] and immunoreactivityfor FGFR1 through FGFR4 in developing and adult rat retina,as well as in adult human retina [53]. The present study con-firms and extends those findings, showing detailed patterns ofFGFR1-IR through FGFR4-IR in central retina of fetal andadult macaque retina, including differential immunoreactiv-ity on distinct components of central cones.

We also show expression of mRNA for FGFR1 throughFGFR4, along with FGFrl1. Those results suggest at least lowlevels of mRNA for each FGF receptor in each cell layer inboth fetal and adult retina with the exception of FGFrl1, whichis not expressed in either the INL or ONL in adult retina, butis intense in the GCL. Distribution of the receptor proteins isnot as widespread as the patterns of mRNA expression mightsuggest. While high levels of immunoreactivity in the GCLmight be expected on the basis of mRNA expression of eachfor the receptors, the findings suggest significant post-tran-scriptional regulation of FGFR1 and FGFR4 protein in theINL and discrete distributions for FGFR1, FGFR3, and FGFR4proteins in outer retina. Most notable of these are the localiza-tion of FGFR3 on horizontal cell processes, cone somas, andthe proximal parts of the fibers of Henle, along with the distri-bution of FGFR4 on almost all parts of the cone.

Inner retina: Previous studies indicate significant rolesfor FGF ligands and receptors in ganglion cell neurite out-growth and survival of ganglion cell populations [39,80-83].In chick retina, blocking FGF receptor activation inhibits gan-glion cell differentiation, while stimulation with FGF1 pro-motes precocious differentiation of ganglion cells [37]. Nospecific analysis has been made of the FGF receptors stimu-lated in those studies. Our findings and those of Kinkl et al.[53] indicate that each of the FGF receptors and FGFrl1(present study) are expressed in the GCL during developmentand in adults and, since all are activated by a number of FGFligands [24], have potential roles in ganglion cell differentia-tion and survival.

Expression of FGFRs in the INL is generally lower thanin either of the other two cell layers. Despite counter-immunolabelling with anti-vimentin or anti-CRALBP, mRNA

expression by and immunoreactivity of Müller cells was noteasy to detect; however, the results suggest that during devel-opment, both FGFR3 and FGFR4 are expressed by Müllercells more intensely than either FGFR1 or FGFR2. WhilemRNA for each of the receptors is detected in presumed neu-rons in the inner and outer parts of the INL in developing andadult retinae, FGFR4-IR is virtually absent and FGFR1-IR islow in the adult INL. The predominant cell types expressingFGFR2 and FGFR3 in the INL appear to be displaced gan-glion cells (present study) and horizontal cells ([53] and presentstudy).

Outer retina: Our finding of low to absent FGFR2-IR onphotoreceptors contrasts with the findings of Kinkl et al. [53].They showed strong immunoreactivity for FGFR2 in the outersegments of cone photoreceptors in frozen sections of adulthuman retina [53]. In this study we detected immunoreactiv-ity to FGFR1 and FGFR3 in a discrete locus on the outer as-pect of the developing inner segments, suggestive of an earlystage of outer segment formation. However, we did not detectimmunoreactivity to any of the FGF receptors in the outersegments of mature photoreceptors. It is not known if this dif-ference is attributable to different approaches to fixation andembedding in the two studies, or if it is due to inter-speciesvariation.

The present results show generally higher levels of im-munoreactivity for FGFR1, FGFR3, and FGFR4 comparedwith FGFR2 in the ONL of fetal and adult retina and indicateonly modest changes in the distribution of receptors betweenthe fetal and adult periods. We did not characterize immu-noreactivity for FGFrl1. During development, FGFR1, FGFR2,FGFR3, and FGFR4 immunoreactivities were detected on pro-cesses that appear to be the RPE microvilli that invest the de-veloping outer segments, with FGFR1-IR the most intense.Unfortunately, in many cases the retinas (along with the pre-sumed microvilli) were detached from the RPE cells and itwas not possible to observe FGFR expression in RPE cell cy-toplasm. In adult retina, however, the presumed RPE microvilliwere not strongly immunoreactive. This may suggest that FGFsignalling is important in establishing the relationship betweenthe RPE microvilli and cone outer segments during develop-ment. Consistent with other reports [84,85], we detectedFGFR-IR in the RPE where FGF signalling appears to have arole in outer segment phagocytosis [84,86]. The results fromour in situ hybridization experiments are difficult to interpretin relation to RPE, due to the tendency of our sense probes tobind non-specifically to the RPE, but not neural retina (Figure2C,F).

FGFR1-IR is more broadly distributed in adult comparedwith fetal retina, however, the distinct and early detection ofFGFR1 at the level of the developing cone pedicles, as well aspersistence of FGFR1-IR in the adult pedicles, suggests a spe-cific role in formation and maintenance of synaptic structures.Cone pedicles also show membranous immunoreactivity toFGFR3 (fetal and adult), with FGFR4 being localized only tothe basal, synaptic region of the pedicle in adults. Postsynap-tically, both FGFR3 and FGFR4 are expressed on OPL pro-cesses, FGFR3 more intensely than FGFR4, with different


10

patterns of immunoreactivity suggesting localization to dif-ferent OPL elements. Given the intense immunoreactivity ofwhat appear to be horizontal cells to anti-FGFR3 antibody, itis tempting to attribute the FGFR3-IR in the OPL to horizon-tal cell processes, although this cannot be stated with certainty.Comparison with descriptions of the organization of cone syn-apses, however, are consistent with this interpretation. In theOPL of the macaque retina, bipolar cell processes aggregatein a narrow band adjacent to the basal portion of the conepedicle, while horizontal cell processes occupy most of thethickness of the OPL, on the inner aspect of the layer of bipo-lar cell processes [87]. This would suggest that the band ofprocesses that are non-reactive to FGFR3 (Figure 5) comprisesbipolar cells that do not express FGFR3. By comparison, sincethe full thickness of the OPL is FGFR4-IR, it appears that theprocesses of both horizontal and bipolar cells express FGFR4(Figure 6).

Significance: Overall, our findings suggest that immu-noreactivities for FGFR3 and FGFR4 are more widespread inthe primate retina and more intense than either FGFR1 orFGFR2. Experimental studies indicate that the neurotrophiceffects of FGFs on photoreceptors are mediated, at least inpart, through upregulation of FGFR1 [73,75,88-90]. Thepresent analysis suggests that expression of FGFR3 andFGFR4 by photoreceptors is constitutively higher than forFGFR1, suggesting that an examination of the roles of FGFR3and FGFR4 in mediation of the neurotrophic effects of FGFson photoreceptors is required.

A significant finding in the present study is the early andintense immunoreactivity of cones to FGFR4. Foveal conephotoreceptors are amongst the first cells generated in the retinaand can be identified as early as Fd 38, while rods in the vicin-ity of the incipient fovea are generated later, from about Fd 45[5]. Cone opsins are detectable by immunolabeling in macaquefovea from around Fd 75, or about 5 days earlier by in situhybridization [9], while other cone-specific markers includ-ing peripherin and α-transducin are detected either simulta-neously with or shortly after opsin expression [91]. Rhodop-sin-IR is detected on the membranes of rods near the incipientfovea from around Fd 65 [92]. In this study, we detected manyintensely FGFR4-IR cones in central retina at Fd 65, suggest-ing an initial expression somewhat earlier, although sectionsfrom younger animals were not available to verify this sug-gestion. We also show co-localization of FGFR4 with α-transducin at Fd 85, confirming the identity of the FGFR4-IRcones. We found no co-localization of rhodopsin and FGFR4in fetal retina at Fd 73 or later, indicating that FGFR4 is anearly, specific marker of cones. We note, however, that FGFR4is expressed in photoreceptors in Xenopus retina, where pre-sumably it is not confined to cones [41]. Specific labelling ofcones for FGFR4 also appears to occur in adult macaque retina,although the localization of rhodopsin to the outer segment ofmature photoreceptors, and of FGFR4 to the soma and axonregions means that a precise assessment of cellular co-local-ization of the two markers is not possible (not shown).

Of particular interest is the role that FGF signalling mayhave in the morphogenesis of foveal cones. The high acuity

function of the foveal region is directly related to a very highlocal density of cone photoreceptors [1] that is established pro-gressively during fetal life and the first years postnatal[12,14,19]. This increase in density takes place in the absenceof cell division [19,93,94] and in association with a decreasein the diameter of cone somata [19] and inner segments [14],suggesting that narrowing and elongation of foveal cones isthe mechanism underlying the crowding of cones into centralretina. Understanding of the cellular mechanisms that governthe morphogenesis of cone photoreceptors, including elabo-ration of the fibers of Henle, is fundamental to understandinghow the central retina becomes specialised for high acuity vi-sion. Our present data show that cones are FGFR4-IR through-out the axon, soma and inner segment from very early in de-velopment. That is, the distribution of FGFR4 is consistentwith a possible role in changing cell shape, including elabora-tion of the fiber of Henle, although these effects remain to beexplored. FGFR1 and FGFR3 have more restricted distribu-tions on cones during development, their arrangements beingconsistent with potential roles in synaptogenesis.

Photoreceptor loss is a feature of normal aging and is pro-nounced in disorders including macular degeneration and avariety of retinal dystrophies [74,95,96]. Data suggest that rodphotoreceptors are more vulnerable to degeneration than cones[97] but that over time cones will degenerate in the absence ofa substantial rod population [98]. Indeed it has been suggestedthat cones derive a diffusible “survival factor” from rods [99]or Müller cells [100], the identity of which is unknown. Inthis context, understanding the factors that promote cone mor-phogenesis and have roles in sustaining normal structure andfunction is crucial to development of preventative or thera-peutic measures, particularly in the areas of transplantation.Characterization of FGFRs expressed by cones that have thepotential to mediate neurotrophic effects [53] (present study),and the identification of signalling cascades that mediate theseeffects is an important initial phase in developing this under-standing.

ACKNOWLEDGEMENTS This work was supported by National Health and MedicalResearch Council (Australia) grant number 153825, the ClaffyFoundation (EC), the Ophthalmic Research Institute of Aus-tralia, and the Sydney Eye Hospital Foundation (RN).

REFERENCES 1. Curcio CA, Sloan KR, Kalina RE, Hendrickson AE. Human pho-

toreceptor topography. J Comp Neurol 1990; 292:497-523.2. Dacey DM. Circuitry for color coding in the primate retina. Proc

Natl Acad Sci U S A 1996; 93:582-8.3. Dacey DM. Primate retina: cell types, circuits and color opponency.

Prog Retin Eye Res 1999; 18:737-63 [published erratum ap-pears in Prog Retin Eye Res 2000; 19:following 646].

4. Dacey DM. Parallel pathways for spectral coding in primate retina.Annu Rev Neurosci 2000; 23:743-75.

5. La Vail MM, Rapaport DH, Rakic P. Cytogenesis in the monkeyretina. J Comp Neurol 1991; 309:86-114.

6. Smelser GK, Ozanics V, Rayborn M, Sagun D. Retinalsynaptogenesis in the primate. Invest Ophthalmol 1974; 13:340-


11

61.7. Linberg KA, Fisher SK. A burst of differentiation in the outer pos-

terior retina of the eleven-week human fetus: an ultrastructuralstudy. Vis Neurosci 1990; 5:43-60.

8. Hendrickson AE. Synaptic development in macaque monkey retinaand its implications for other developmental sequences. PerspectDev Neurobiol 1996; 3:195-201.

9. Bumsted K, Jasoni C, Szel A, Hendrickson A. Spatial and tempo-ral expression of cone opsins during monkey retinal develop-ment [published erratum appears in J Comp Neurol 1997;380:291]. J Comp Neurol 1997; 378:117-34.

10. Hendrickson AE. Development of the primate retina. In: MeisamiE, Timiras PS, editors. Handbook of human growth and devel-opmental biology. Boca Raton (FL): CRC Press; 1988. p. 165-78.

11. Hendrickson A. A morphological comparison of foveal develop-ment in man and monkey. Eye 1992; 6:136-44.

12. Hendrickson A. Primate foveal development: a microcosm ofcurrent questions in neurobiology. Invest Ophthalmol Vis Sci1994; 35:3129-33.

13. Hendrickson AE, Yuodelis C. The morphological developmentof the human fovea. Ophthalmology 1984; 91:603-12.

14. Yuodelis C, Hendrickson A. A qualitative and quantitative analy-sis of the human fovea during development. Vision Res 1986;26:847-55.

15. Provis JM, Diaz CM, Dreher B. Ontogeny of the primate fovea:a central issue in retinal development. Prog Neurobiol 1998;54:549-80.

16. Provis JM, Sandercoe T, Hendrickson AE. Astrocytes and bloodvessels define the foveal rim during primate retinal develop-ment. Invest Ophthalmol Vis Sci 2000; 41:2827-36.

17. Packer O, Hendrickson AE, Curcio CA. Photoreceptor topogra-phy of the retina in the adult pigtail macaque (Macacanemestrina). J Comp Neurol 1989; 288:165-83.

18. Packer O, Hendrickson AE, Curcio CA. Development redistribu-tion of photoreceptors across the Macaca nemestrina (pigtailmacaque) retina. J Comp Neurol 1990; 298:472-93.

19. Diaz-Araya C, Provis JM. Evidence of photoreceptor migrationduring early foveal development: a quantitative analysis of hu-man fetal retinae. Vis Neurosci 1992; 8:505-14.

20. Ornitz DM. FGFs, heparan sulfate and FGFRs: complex interac-tions essential for development. Bioessays 2000; 22:108-12.

21. Ornitz DM, Itoh N. Fibroblast growth factors. Genome Biol 2001;2:REVIEWS3005.

22. Johnson DE, Williams LT. Structural and functional diversity inthe FGF receptor multigene family. Adv Cancer Res 1993; 60:1-41.

23. Partanen J, Vainikka S, Korhonen J, Armstrong E, Alitalo K. Di-verse receptors for fibroblast growth factors. Prog Growth Fac-tor Res 1992; 4:69-83.

24. Ornitz DM, Xu J, Colvin JS, McEwen DG, MacArthur CA,Coulier F, Gao G, Goldfarb M. Receptor specificity of the fi-broblast growth factor family. J Biol Chem 1996; 271:15292-7.

25. Sleeman M, Fraser J, McDonald M, Yuan S, White D, GrandisonP, Kumble K, Watson JD, Murison JG. Identification of a newfibroblast growth factor receptor, FGFR5. Gene 2001; 271:171-82.

26. Kim I, Moon S, Yu K, Kim U, Koh GY. A novel fibroblast growthfactor receptor-5 preferentially expressed in the pancreas(1).Biochim Biophys Acta 2001; 1518:152-6.

27. Wiedemann M, Trueb B. The mouse FGFrl1 gene coding for anovel FGF receptor-like protein. Biochim Biophys Acta 2001;1520:247-50.

28. Leelayuwat C, Abraham LJ, Pinelli M, Townend DC, Wilks AF,Dawkins RL. The primate MHC contains sequences related tothe fibroblast growth factor receptor gene family. Tissue Anti-gens 1996; 48:59-64.

29. Ford-Perriss M, Abud H, Murphy M. Fibroblast growth factorsin the developing central nervous system. Clin Exp PharmacolPhysiol 2001; 28:493-503.

30. Murphy M, Drago J, Bartlett PF. Fibroblast growth factor stimu-lates the proliferation and differentiation of neural precursor cellsin vitro. J Neurosci Res 1990; 25:463-75.

31. Lewis GP, Erickson PA, Guerin CJ, Anderson DH, Fisher SK.Basic fibroblast growth factor: a potential regulator of prolif-eration and intermediate filament expression in the retina. JNeurosci 1992; 12:3968-78.

32. Bartlett PF, Brooker GJ, Faux CH, Dutton R, Murphy M, TurnleyA, Kilpatrick TJ. Regulation of neural stem cell differentiationin the forebrain. Immunol Cell Biol 1998; 76:414-8.

33. Bailly K, Soulet F, Leroy D, Amalric F, Bouche G. Uncouplingof cell proliferation and differentiation activities of basic fibro-blast growth factor. FASEB J 2000; 14:333-44.

34. Desire L, Head MW, Fayein NA, Courtois Y, Jeanny JC. Sup-pression of fibroblast growth factor 2 expression by antisenseoligonucleotides inhibits embryonic chick neural retina cell dif-ferentiation and survival in vivo. Dev Dyn 1998; 212:63-74.

35. Hajihosseini MK, Dickson C. A subset of fibroblast growth fac-tors (Fgfs) promote survival, but Fgf-8b specifically promotesastroglial differentiation of rat cortical precursor cells. Mol CellNeurosci 1999; 14:468-85.

36. Desire L, Courtois Y, Jeanny JC. Endogenous and exogenousfibroblast growth factor 2 support survival of chick retinal neu-rons by control of neuronal neuronal bcl-x(L) and bcl-2 expres-sion through a fibroblast berowth factor receptor 1- and ERK-dependent pathway. J Neurochem 2000; 75:151-63.

37. McCabe KL, Gunther EC, Reh TA. The development of the pat-tern of retinal ganglion cells in the chick retina: mechanismsthat control differentiation. Development 1999; 126:5713-24.

38. Brittis PA, Silver J, Walsh FS, Doherty P. Fibroblast growth fac-tor receptor function is required for the orderly projection ofganglion cell axons in the developing mammalian retina. MolCell Neurosci 1996; 8:120-8.

39. Walz A, McFarlane S, Brickman YG, Nurcombe V, Bartlett PF,Holt CE. Essential role of heparan sulfates in axon navigationand targeting in the developing visual system. Development1997; 124:2421-30.

40. Deng CX, Wynshaw-Boris A, Shen MM, Daugherty C, OrnitzDM, Leder P. Murine FGFR-1 is required for earlypostimplantation growth and axial organization. Genes Dev1994; 8:3045-57.

41. McFarlane S, Zuber ME, Holt CE. A role for the fibroblast growthfactor receptor in cell fate decisions in the developing verte-brate retina. Development 1998; 125:3967-75.

42. Arman E, Haffner-Krausz R, Chen Y, Heath JK, Lonai P. Tar-geted disruption of fibroblast growth factor (FGF) receptor 2suggests a role for FGF signaling in pregastrulation mamma-lian development. Proc Natl Acad Sci U S A 1998; 95:5082-7.

43. Raffioni S, Thomas D, Foehr ED, Thompson LM, Bradshaw RA.Comparison of the intracellular signaling responses by threechimeric fibroblast growth factor receptors in PC12 cells. ProcNatl Acad Sci U S A 1999; 96:7178-83.

44. Heuer JG, von Bartheld CS, Kinoshita Y, Evers PC, Bothwell M.Alternating phases of FGF receptor and NGF receptor expres-sion in the developing chicken nervous system. Neuron 1990;5:283-96.


12

45. Ohuchi H, Koyama E, Myokai F, Nohno T, Shiraga F, Matsuo T,Matsuo N, Taniguchi S, Noji S. Expression patterns of two fi-broblast growth factor receptor genes during early chick eyedevelopment. Exp Eye Res 1994; 58:649-58.

46. Tcheng M, Fuhrmann G, Hartmann MP, Courtois Y, Jeanny JC.Spatial and temporal expression patterns of FGF receptor genestype 1 and type 2 in the developing chick retina. Exp Eye Res1994; 58:351-8.

47. Fayein NA, Head MW, Jeanny JC, Courtois Y, Fuhrmann G.Expression of the chicken cysteine-rich fibroblast growth fac-tor receptor (CFR) during embryogenesis and retina develop-ment. J Neurosci Res 1996; 43:602-12.

48. Torriglia A, Jeanny JC, Blanquet PR. Immunohistochemical analy-sis of fibroblast growth factor receptor in bovine retina. NeurosciLett 1994; 172:125-8.

49. Torriglia A, Blanquet PR. Immunochemical evidence for a fibro-blast growth factor receptor in adult retinal optic fiber and syn-aptic layers. Neuroscience 1994; 60:969-81.

50. Blanquet PR, Jonet L. Signal-regulated proteins and fibroblastgrowth factor receptors: comparative immunolocalization in ratretina. Neurosci Lett 1996; 214:135-8.

51. de Iongh R, Lovicu FJ, Chamberlain CG, McAvoy JW. Differen-tial expression of fibroblast growth factor receptors during ratlens morphogenesis and growth. Invest Ophthalmol Vis Sci1997; 38:1688-99.

52. Fuhrmann V, Kinkl N, Leveillard T, Sahel J, Hicks D. Fibroblastgrowth factor receptor 4 (FGFR4) is expressed in adult rat andhuman retinal photoreceptors and neurons. J Mol Neurosci 1999;13:187-97.

53. Kinkl N, Hageman GS, Sahel JA, Hicks D. Fibroblast growthfactor receptor (FGFR) and candidate signaling molecule dis-tribution within rat and human retina. Mol Vis 2002; 8:149-60 .

54. Kitaoka T, Aotaki-Keen AE, Hjelmeland LM. Distribution of FGF-5 in the rhesus macaque retina [published erratum appears inInvest Ophthalmol 1995; 36:976]. Invest Ophthalmol Vis Sci1994; 35:3189-98.

55. Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVailMM. Photoreceptor degeneration in inherited retinal dystrophydelayed by basic fibroblast growth factor. Nature 1990; 347:83-6.

56. Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVailMM. Basic fibroblast growth factor and local injury protectphotoreceptors from light damage in the rat. J Neurosci 1992;12:3554-67.

57. LaVail MM, Yasumura D, Matthes MT, Lau-Villacorta C, UnokiK, Sung CH, Steinberg RH. Protection of mouse photorecep-tors by survival factors in retinal degenerations. InvestOphthalmol Vis Sci 1998; 39:592-602.

58. Xiao M, Sastry SM, Li ZY, Possin DE, Chang JH, Klock IB,Milam AH. Effects of retinal laser photocoagulation on photo-receptor basic fibroblast growth factor and survival. InvestOphthalmol Vis Sci 1998; 39:618-30.

59. McAvoy JW, Chamberlain CG. Fibroblast growth factor (FGF)induces different responses in lens epithelial cells depending onits concentration. Development 1989; 107:221-8.

60. McAvoy JW, Chamberlain CG, de Iongh RU, Richardson NA,Lovicu FJ. The role of fibroblast growth factor in eye lens de-velopment. Ann N Y Acad Sci 1991; 638:256-74.

61. Gu X, Seong GJ, Lee YG, Kay EP. Fibroblast growth factor 2uses distinct signaling pathways for cell proliferation and cellshape changes in corneal endothelial cells. Invest OphthalmolVis Sci 1996; 37:2326-34.

62. de Iongh RU, Lovicu FJ, Hanneken A, Baird A, McAvoy JW.

FGF receptor-1 (flg) expression is correlated with fibre differ-entiation during rat lens morphogenesis and growth. Dev Dyn1996; 206:412-26.

63. Kay EP, Park SY, Ko MK, Lee SC. Fibroblast growth factor 2uses PLC-gamma1 for cell proliferation and PI3-kinase for al-teration of cell shape and cell proliferation in corneal endothe-lial cells. Mol Vis 1998; 4:22 .

64. Coulier F, Pontarotti P, Roubin R, Hartung H, Goldfarb M,Birnbaum, D. Of worms and men: An evolutionary perspectiveon fibroblast growth factor (FGF) and FGF receptor families. JMol Evol 1997; 44:43-56.

65. Lerea CL, Bunt-Milam AH, Hurley JB. Alpha-transducin ispresent in blue-, green-, and red-sensitive cone photoreceptorsin the human retina. Neuron 1989; 3:367-76.

66. Bumsted-O’Brien K, Schulte, D Hendrickson A. Expression ofphotoreceptor-associated molecules during human fetal eye de-velopment. Mol Vis 2003; 9:401-409.

67. Caruelle D, Groux-Muscatelli B, Gaudric A, Sestier C, CoscasG, Caruelle JP, Barritault D. Immunological study of acidic fi-broblast growth factor (aFGF) distribution in the eye. J CellBiochem 1989; 39:117-28.

68. Baudouin C, Fredj-Reygrobellet D, Caruelle JP, Barritault D,Gastaud P, Lapalus P. Acidic fibroblast growth factor distribu-tion in normal human eye and possible implications in ocularpathogenesis. Ophthalmic Res 1990; 22:73-81.

69. Noji S, Matsuo T, Koyama E, Yamaai T, Nohno T, Matsuo N,Taniguchi S. Expression pattern of acidic and basic fibroblastgrowth factor genes in adult rat eyes. Biochem Biophys ResCommun 1990; 168:343-9.

70. Jacquemin E, Jonet L, Oliver L, Bugra K, Laurent M, Courtois Y,Jeanny JC. Developmental regulation of acidic fibroblast growthfactor (aFGF) expression in bovine retina. Int J Dev Biol 1993;37:417-23.

71. Bugra K, Oliver L, Jacquemin E, Laurent M, Courtois Y, HicksD. Acidic fibroblast growth factor is expressed abundantly byphotoreceptors within the developing and mature rat retina [pub-lished erratum appears in Eur J Neurosci 1994; 6:1062]. Eur JNeurosci 1993; 5:1586-95.

72. Wen R, Song Y, Cheng T, Matthes MT, Yasumura D, LaVail MM,Steinberg RH. Injury-induced upregulation of bFGF and CNTFmRNAS in the rat retina. J Neurosci 1995; 15:7377-85.

73. Ozaki S, Radeke MJ, Anderson DH. Rapid upregulation of fibro-blast growth factor receptor 1 (flg) by rat photoreceptor cellsafter injury. Invest Ophthalmol Vis Sci 2000; 41:568-79.

74. Gao H, Hollyfield JG. Basic fibroblast growth factor in retinaldevelopment: differential levels of bFGF expression and con-tent in normal and retinal degeneration (rd) mutant mice. DevBiol 1995; 169:168-84.

75. Yamamoto C, Ogata N, Matsushima M, Takahashi K, MiyashiroM, Yamada H, Maeda H, Uyama M, Matsuzaki K. Gene ex-pressions of basic fibroblast growth factor and its receptor inhealing of rat retina after laser photocoagulation. Jpn JOphthalmol 1996; 40:480-90.

76. Yamamoto C, Ogata N, Yi X, Takahashi K, Miyashiro M, YamadaH, Uyama M, Matsuzaki K. Immunolocalization of basic fibro-blast growth factor during wound repair in rat retina after laserphotocoagulation. Graefes Arch Clin Exp Ophthalmol 1996;234:695-702.

77. Hackett SF, Schoenfeld CL, Freund J, Gottsch JD, Bhargave S,Campochiaro PA. Neurotrophic factors, cytokines and stressincrease expression of basic fibroblast growth factor in retinalpigmented epithelial cells. Exp Eye Res 1997; 64:865-73.

78. Valter K, Maslim J, Bowers F, Stone J. Photoreceptor dystrophy


13

in the RCS rat: roles of oxygen, debris, and bFGF. InvestOphthalmol Vis Sci 1998; 39:2427-42.

79. Xiao M, McLeod D, Cranley J, Williams G, Boulton M. Growthfactor staining patterns in the pig retina following retinal laserphotocoagulation. Br J Ophthalmol 1999; 83:728-36.

80. Sievers J, Hausmann B, Unsicker K, Berry M. Fibroblast growthfactors promote the survival of adult rat retinal ganglion cellsafter transection of the optic nerve. Neurosci Lett 1987; 76:157-62.

81. Lipton SA, Wagner JA, Madison RD, D’Amore PA. Acidic fi-broblast growth factor enhances regeneration of processes bypostnatal mammalian retinal ganglion cells in culture. Proc NatlAcad Sci U S A 1988; 85:2388-92.

82. Bahr M, Vanselow J, Thanos S. Ability of adult rat ganglion cellsto regrow axons in vitro can be influenced by fibroblast growthfactor and gangliosides. Neurosci Lett 1989; 96:197-201.

83. Chai L, Morris JE. Heparan sulfate in the inner limiting mem-brane of embryonic chicken retina binds basic fibroblast growthfactor to promote axonal outgrowth. Exp Neurol 1999; 160:175-85.

84. Malecaze F, Mascarelli F, Bugra K, Fuhrmann G, Courtois Y,Hicks D. Fibroblast growth factor receptor deficiency in dys-trophic retinal pigmented epithelium. J Cell Physiol 1993;154:631-42.

85. Alizadeh M, Miyamura N, Handa JT, Hjelmeland LM. HumanRPE cells express the FGFR2IIIc and FGFR3IIIc splice vari-ants and FGF9 as a potential high affinity ligand. Exp Eye Res2003; 76:249-56.

86. McLaren MJ, Inana G. Inherited retinal degeneration: basic FGFinduces phagocytic competence in cultured RPE cells from RCSrats. FEBS Lett 1997; 412:21-9.

87. Haverkamp S, Grunert U, Wassle H. The cone pedicle, a com-plex synapse in the retina. Neuron 2000; 27:85-95.

88. Fontaine V, Kinkl N, Sahel J, Dreyfus H, Hicks D. Survival ofpurified rat photoreceptors in vitro is stimulated directly by fi-broblast growth factor-2. J Neurosci 1998; 18:9662-72.

89. Guillonneau X, Regnier-Ricard F, Laplace O, Jonet L, BryckaertM, Courtois Y, Mascarelli F. Fibroblast growth factor (FGF)soluble receptor 1 acts as a natural inhibitor of FGF2 neu-

rotrophic activity during retinal degeneration. Mol Biol Cell1998; 9:2785-802.

90. Geller SF, Lewis GP, Fisher SK. FGFR1, signaling, and AP-1expression after retinal detachment: reactive Muller and RPEcells. Invest Ophthalmol Vis Sci 2001; 42:1363-9.

91. Sears S, Erickson A, Hendrickson A. The spatial and temporalexpression of outer segment proteins during development ofMacaca monkey cones. Invest Ophthalmol Vis Sci 2000; 41:971-9.

92. Dorn EM, Hendrickson L, Hendrickson AE. The appearance ofrod opsin during monkey retinal development. InvestOphthalmol Vis Sci 1995; 36:2634-51.

93. Provis JM, van Driel D. Retinal development in humans: the rolesof differential growth rates, cell migration and naturally occur-ring cell death. Aust N Z J Ophthalmol 1985; 13:125-33.

94. Sandercoe TM, Madigan MC, Billson FA, Penfold PL, ProvisJM. Astrocyte proliferation during development of the humanretinal vasculature. Exp Eye Res 1999; 69:511-23.

95. Frank RN. Growth factors in age-related macular degeneration:pathogenic and therapeutic implications. Ophthalmic Res 1997;29:341-53.

96. Li ZY, Chang JH, Milam AH. Distribution of basic fibroblastgrowth factor in human retinas with retinitis pigmentosa. ExpEye Res 1997; 65:855-9.

97. Hicks D, Sahel J. The implications of rod-dependent cone sur-vival for basic and clinical research. Invest Ophthalmol Vis Sci1999; 40:3071-4.

98. Curcio CA, Owsley C, Jackson GR. Spare the rods, save the conesin aging and age-related maculopathy. Invest Ophthalmol VisSci 2000; 41:2015-8.

99. Mohand-Said S, Deudon-Combe A, Hicks D, Simonutti M, ForsterV, Fintz AC, Leveillard T, Dreyfus H, Sahel JA. Normal retinareleases a diffusible factor stimulating cone survival in the reti-nal degeneration mouse. Proc Natl Acad Sci U S A 1998;95:8357-62.

100. Wahlin KJ, Campochiaro PA, Zack DJ, Adler R. Neurotrophicfactors cause activation of intracellular signaling pathways inMuller cells and other cells of the inner retina, but not photore-ceptors. Invest Ophthalmol Vis Sci 2000; 41:927-36.


14

The print version of this article was created on 8 Jan 2004. This reflects all typographical corrections and errata to the article through that date.Details of any changes may be found in the online version of the article.

Documents

Differential distribution of fibroblast growth factor ... · The pedicles of developing and adult cones are FGFR1 and FGFR3 immunoreactive, and the basal, synaptic region is FGFR4