Basic fibroblast growth factor applied to the optic nerve after injury increases long‐term cell survival in the frog retina

Basic fibroblast growth factor applied to the optic nerve after injury increases long‐term cell survival in the frog retina

Please download to get full document.

View again

of 13
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 14 | Pages: 13

Extension: PDF | Download: 0

  Basic Fibroblast Growth Factor Appliedto the Optic Nerve After Injury Increases Long-Term Cell Survival in theFrog Retina ROSA E. BLANCO,* ARGELIO LO´PEZ-ROCA, JORGE SOTO,  AND JONATHAN M. BLAGBURN Institute of Neurobiology and Departments of Anatomy and Physiology, University of Puerto Rico Medical Sciences Campus, San Juan, Puerto Rico 00901 ABSTRACTThe neuroprotective effects of basic fibroblast growth factor (bFGF) on the long-termsurvival of axotomized retinal ganglion cells (RGCs) were studied in the frog   Rana pipiens .Cell loss was quantified in different regions of the ganglion cell layer using Nissl staining andtetramethylrhodamine dextran amine backfilling. All regions of the retina showed a signif-icant decrease (32–66%) in RGC numbers between 4 and 16 weeks after axotomy. Some cellsshowed morphological and biochemical signs of apoptosis. A single application of bFGF to theoptic nerve stump at the time of axotomy protected many of the cells 6 weeks after the injury,but this effect was lost by 12 weeks. A second application of bFGF, 6 weeks after the injury,rescued many RGCs at 12 weeks. In contrast, single or double injections of bFGF into theeyeball had no effect on RGC survival. Axotomized RGCs were significantly enlarged andelongated after axotomy, and these morphological changes were increased by bFGF treat-ment. In the normal retina and optic nerve, immunocytochemical staining showed bFGF-likeimmunoreactivity (-LI) in the pigment epithelial layer, in the outer segments of photorecep-tors, and in occasional RGCs. Strong bFGF-LI was present in Mu¨ller cells and in optic nerveastrocytes and oligodendrocytes. FGF receptor-LI was present in photoreceptors, outer plex-iform layer, retinal ganglion cell axons, and Mu¨ller cells. FGF receptor-LI was also observedin optic nerve glia. J. Comp. Neurol. 423:646–658, 2000.  ©  2000 Wiley-Liss, Inc. Indexing terms: growth factor; regeneration; astrocyte; apoptosis; cell death  Axotomized mammalian retinal ganglion cells (RGCs)die rapidly, in large numbers (Miller and Oberdorfer,1981; Grafstein and Ingoglia, 1982). Although some maydie immediately from necrosis following axonal injury(Bien et al., 1999), the death of many RGCs in the longerterm is probably due to disruption of the supply of trophicfactors from the target tissues (Barde, 1989; Raff et al.,1993; for review see Pettmann and Henderson, 1998). Thisappears to result in programmed cell death, or apoptosis(Garcı´a-Valenzuela et al., 1994; Isenmann et al., 1997). Although overexpression of the apoptosis-preventing pro-tein Bcl-2 largely prevents cell death for a short time afteraxotomy of mouse optic nerve, 30% cell death still occursat longer intervals (Chierzi et al., 1998), so the mecha-nisms by which this delayed cell death occurs remainunclear. As in mammals, a large proportion (approximately 40–70%) of axotomized frog RGCs also die, often many weeksafter the injury (Scalia et al., 1985). The mechanisms bywhich these cells die are not known. This long-term celldeath takes place even though frog RGC axons are knownto regenerate successfully, allowing recovery of vision(Sperry, 1944). The optic nerve glia of lower vertebrates donot exhibit the inhibitory properties of optic nerve glia inmammals (Wanner et al., 1995; Ankerhold et al., 1998),which probably accounts for the different regenerativeabilities of RGCs in those groups. The frog optic system istherefore a useful experimental model with which to in- vestigate how the application of neurotrophic factors tothe nerve modulates the long-term survival and regener- *Correspondence to: R.E. Blanco, PhD, Institute of Neurobiology, 201Boulevard del Valle, San Juan, PR 00901.E-mail: rblanco@neurobio.upr.clu.eduReceived 14 July 1999; Revised 28 March 2000; Accepted 29 March 2000 THE JOURNAL OF COMPARATIVE NEUROLOGY 423:646–658 (2000) ©  2000 WILEY-LISS, INC.  ation of RGCs. It has the additional advantage that, un-like the case with mammalian optic nerve, a large portionof the frog optic nerve lies extracranially and is thus veryaccessible to experimental manipulation.Basic fibroblast growth factor (bFGF, also termedFGF-2) has been isolated in relatively high concentrationsfrom the brain and retina (D’Amore and Klagsbrun, 1984;Baird et al., 1995), and, in the visual system, it has beenimplicated as a trophic factor that protects against injuryand degeneration and stimulates regeneration (Hicks,1996; Viollet and Doherty, 1997). A single application of bFGF affords long-term protection to axotomized lateralgeniculate neurons (Agarwala and Kalil, 1998). In vitrobFGF prolongs survival of RGCs in explants and stimu-lates differentiation and survival of photoreceptors (Ba¨hret al., 1989; Hicks and Courtois, 1992; Fontaine et al.,1998). The effects of bFGF on modulating survival of axo-tomized RGCs are not clear. Although application of bFGFto the transected optic nerves of rats increases the sur- vival of ganglion cells 1 month after axotomy (Sievers etal., 1987), long-term survival effects have not been re-ported. Interpretation of these results is complicated bythe fact that bFGF apparently has little neuroprotectiveeffect when applied intraocularly after optic nerve crush(Schmitt and Sabel, 1996).The objective of the present study was to determine thetime course of cell death in the frog retina after axotomy,to investigate whether apoptosis is occurring, and to ex-plore whether bFGF application can rescue ganglion cellsin this system. In addition, immunocytochemistry wasused to localize endogenous bFGF and the receptor ty-rosinekinaseFGFR1,whichhasbeenshowntobindbFGFwith high affinity (Jaye et al., 1992). Some of these datahave been presented previously in abstract form (Blancoet al., 1997). MATERIALS AND METHODS Animals  Adult  Rana pipiens  were obtained from commercialsources and kept in aquaria in recirculating tap water at16°C. The animals were fed live crickets twice per week. All animals used for experiments were 7–8 cm in length. At least four to six animals were used for each experimen-tal treatment. All protocols conformed to NIH guidelinesand were approved by the institutional animal care anduse committee. Cutting the optic nerve Theanimalwasanesthetizedbyimmersioninasolutionof 0.3% tricaine, then a slit was cut through the palate, themuscles were separated, and the optic nerve was cut orcrushed. After an overnight recovery period the animalwas replaced in the colony. The survival rate of the oper-ated frogs was almost 100%.  Application of growth factor bFGF (Boehringer-Mannheim; Indianapolis, IN) wasdissolved at a concentration of 10   g/ml in phosphate-buffered saline (PBS), pH 7.4. In some experiments 5   lbFGF solution were injected intraocularly with a Hamil-ton syringe; in others the bFGF solution was applied di-rectly to the optic nerve stump, to give a total dose of 50 ng bFGF. At the time of cutting, the nerve stump was liftedand placed on a strip of Parafilm, and bFGF solution wasapplied to the tip for 10 min. The Parafilm was removedand the animal sutured. In some animals, the procedurewas repeated 6 weeks after cutting the nerve. Controlapplications consisted of 5  l PBS. Counts of cells in the ganglion cell layer The frog was anesthetized and the heart exposed forperfusion with frog Ringer solution, followed by fixative.The latter consisted of 1% glutaraldehyde, 1% paraformal-dehyde in 0.1 M phosphate buffer (pH 7.3). The eye wasthen removed, and the anterior portion cut around the oraserrata and removed. The retina was dissected from thesclera and choroid and the pigmented epithelium carefullyremoved with a fine paintbrush. Tissue was postfixed for2–4 hours, then washed in phosphate buffer. The retinawas flattened onto a gelatinized microscope slide betweena Teflon sheet and a second slide, then air dried overnight. After staining with 0.5% cresyl violet acetate (Nissl stain)for 10 minutes, the retina was dehydrated and mounted inCanada balsam. All retinas used for counting had totalareas of 70–80 mm 2 .Cell counts in the ganglion cell layer (GCL) were madein eight regions, chosen after reference to the data of Humphrey (1987; see Fig. 2). These regions were locatedin the superior, inferior, temporal, and nasal quadrants, 1mm from the optic nerve and two-thirds the distance tothe periphery. Because outward displacement of 5% of RGCs occurs after axotomy (Scalia et al., 1985; Singmanand Scalia, 1990), we counted cells in the major focal planeof the ganglion cell layer and up to 10   m towards theinner nuclear layer. Nissl-stained cells in this range of focus were counted using a 0.1  0.1 mm graticule with a  100 objective, taking the average of three 0.01 mm 2 sample areas for each of the eight regions of the retina.Scattered, small, densely stained cells at the vitreal sur-face were not included in the counts. In situ apoptosis detection The Oncor ApopTag Plus Kit was used to stain cells atthe light microscopy level. This protocol utilizes certainreagents for nonisotopic DNA end-extension in situ andother reagents for immunohistochemical staining of theextended DNA. Residues of digoxigenin-nucleotide arecatalytically added to the DNA by terminal deoxynucleo-tidyl transferase (TdT), an enzyme that catalyzes atemplate-independent addition of nucleotide triphosphateto the 3  -OH ends of double- or single-stranded DNA. Theincorporated nucleotides form a random heteropolymer, ina ratio that has been optimized for antidigoxigenin anti-body binding. The antidigoxigenin antibody fragment isfluorescein-coupled and was visualized with epifluores-cence. TDA labeling of RGCs  At 12 weeks after axotomy, the optic nerve was exposedand several (10–20) crystals of tetramethylrhodaminedextran amine (3,000 MW; Molecular Probes, Eugene,OR) were inserted with a needle into the stump, and thepalate was sutured. After 48 hours, the frog was anesthe-tized and the heart exposed for perfusion with frog Ringersolution, then 4% paraformaldehyde for 30 minutes. Theeyewasthenremoved,theretinadissectedoutandplaced,axon layer uppermost, on a microscope slide, covered witha drop of Vectashield (Vector Laboratories, Burlingame, 647bFGF RESCUES FROG RETINAL GANGLION CELLS  CA), and coverslipped. With these parameters, TDA con-sistently labeled 60–75% of the cells in the GCL. Longerexposure times to TDA stained more ganglion cells, butthey were less readily distinguished from macrophagesbecause of the granular nature of the labeling. LabeledRGCs were quantified by using the same sampling tech-niques as described for Nissl-stained retinas. Quantification of ganglion cell size The outlines of cells in the GCL of Nissl-stained retinaswere traced using a camera lucida. For each treatment(control, 12 weeks cut, and 12 weeks cut    two bFGFapplications to the optic nerve stump), four retinas wereanalyzed by tracing three fields of view, approximately160 ı`m in diameter, from the temporal regions. The areasand dimensions of the cell bodies were measured using Metamorph software (Universal Imaging, Inc.). Immunocytochemistry  Retinas were fixed in 2% paraformaldehyde in 0.1 MPBS for 20 min. After washing in PBS they were soaked in30% sucrose in PBS overnight. After freezing, 10–12   msections were cut from peripheral and central areas. After two 5 min washes, the sections were incubated in10% normal goat serum, then in primary antibody dilutedin 0.1 M PBS  0.3% Triton X-100  0.5 % BSA, for 1 hourat room temperature. The Calbiochem anti-bFGF poly-clonal was used at a dilution of 1:250, whereas the  Xeno- pus  antibody was used at a dilution of 1:750. The rabbitanti-FGFR1 polyclonal was used at a dilution of 1:250. After three 10 min washes in PBS, the sections wereincubated with goat anti-rabbit secondary antibody for 1hour at room temperature. The sections were washed sixtimes for 10 min each in PBS then mounted in Polymount.Preparations were viewed with a Zeiss Axioskop micro-scope equipped with a MicroMAX CCD camera (PrincetonInstruments, Inc., Trenton, NJ). Images were adjusted foroptimal brightness and contrast, and plates were assem-bled using Adobe Photoshop (Adobe Systems, Inc., SanJose, CA).The Ab-2 polyclonal antibody from Calbiochem gave thestrongest specific staining. This is raised against a highlyconserved sequence from human bFGF, which shows 75%identity with the  Xenopus  and fish sequences. A secondantibody raised against recombinant  Xenopus  bFGF (Song and Slack, 1994) gave identical results. This antibody hasbeen shown to be specific for bFGF and does not react withaFGF or eFGF (Song and Slack, 1994). Antibody specific-ity was also tested by omitting the primary antibody andby preabsorbing the antibody with 1   M human bFGF(Alomone Labs, Inc). Both procedures resulted in the ab-sence of all immunostaining.We also used three antibodies directed against the FGFreceptor. The most specific staining was obtained with anaffinity-purified polyclonal antibody raised against  Xeno- pus  FGFR1 (kindly donated by Dr. T. Musci; Musci et al.,1990). Similar results were obtained with the anti-humanFGFR1 (Flg) polyclonal antibody from Santa Cruz Bio-technology (Santa Cruz, CA).Other monoclonal antibodies used were Olig, which isspecific for amphibian oligodendrocytes (Steen et al.,1989), and 40E-C, which labels radial glial cells (Alvarez-Buylla et al., 1987). Both were obtained from the Devel-opmental Studies Hybridoma Bank (University of Iowa). A monoclonal antibody against 160 kD neurofilament pro-tein was obtained from Zymed (San Francisco, CA). RESULTSRGC death after axotomy  Nissl staining of the retina showed a striking decreasein the number of cells in the GCL 26 weeks after axotomy(Fig. 1A,B). At 26 weeks there appeared to be a higherproportion of larger cells than in the control, and many of the cells had a bipolar morphology (Fig. 1B). From 4 to 16weeks after axotomy, some cells (approximately 1–3/0.01mm 2 ) contained what appeared to be apoptotic bodies,evidence of programmed cell death (Fig. 1C). ApopTag staining of transverse sections confirmed that apoptosiswas taking place during this period (Fig. 1D).Counts were made of cells in the GCL of Nissl-stainedretinas at different times after cutting or crushing theoptic nerve. Previous work has shown that the total num-ber of cells in the frog retina is significantly decreasedafter axotomy (Humphrey, 1987), but suggests that differ-ent areas undergo different degrees of cell loss. Based onHumphrey’s cell density plots, eight areas were chosen forcell counts, four near the optic nerve and four in periph-eral regions of the retina, with three 0.01 mm 2 samplestaken from each (Fig. 2). The results are shown in Figure3. In all retinal areas except the peripheral superior andinferior, a significant 25–30% decline in cell numbers wasobserved by 4 weeks after axotomy (Fig. 3). Cell loss con-tinued until about 12 weeks, when these areas had lostapproximately 50% of their population. A 30–40% loss of cells in the peripheral superior and inferior regions wasalso apparent by 12 weeks. In all areas, cell loss continuedat a slower rate until 26 weeks, the last time point stud-ied. Crushing the nerve instead of cutting it consistentlygave significantly greater cell survival rates only at 26weeks; the number of cells surviving at 12 weeks wasgenerally similar in cut and crush preparations (Fig. 3).It is clear from Figure 3 that different percentages of cells died in different regions of the retina. These differ-ences in death rate appear to be proportional to the initialcell density. Thus, in the most densely populated region,the peripheral temporal, 66% of cells died, whereas, in theleast densely populated region, the peripheral superiorretina, only 32% died. What is noticeable is that, whereasthe initial cell densities varied from 60 to 150 cells per0.01 mm 2 , at 26 weeks after cutting the nerve all eightretinal areas had a cell density of about 40–50 cells per0.01 mm 2 .In some animals a segment of the optic nerve was re-moved, and then any outgrowth from the stump was againremoved at 12 weeks, preventing reconnection with thetarget and ensuring maximal cell death of RGCs. Whenexamined 40 weeks after the first cut, the few ganglioncells that remained had large, lightly staining cell bodies(data not shown). The great majority of cells remaining inthe GCL had small, darkly stained nuclei and probablyrepresented displaced amacrine cells. These were uni-formly distributed with a density of 8–11 cells per 0.01mm 2 in all areas, similar to the numbers described byScalia et al. (1985). 648 R.E. BLANCO ET AL.  bFGF applied to the optic nerve stumprescues RGCs  Application of bFGF to the optic nerve stump.  Ap-plication of bFGF (50 ng) to the cut surface of the proximalsegment of the optic nerve immediately after axotomyresulted in significant increases in the numbers of cellssurviving at 6 weeks in all regions of the retina (Fig. 4A),so that 34–80% of the cells that would have died wererescued.The addition of a second bFGF application to the opticnerve at 6 weeks significantly increased the number of cells counted at 12 weeks, in all regions of the retina(Fig. 4A). The percentage of cells saved was more vari-able, ranging from 25% of those that would have died(peripheral temporal region) to 100% (peripheral infe-rior). Two control injections of PBS into the optic nervestump had no significant effects on GCL cell numbers(Fig. 4A).  Application of bFGF intraocularly.  A single intraoc-ular injection of the same amount of bFGF had no signif-icant effects on cell numbers (Fig. 4B). The addition of asecond intraocular injection at 6 weeks also had no signif- Fig. 1. Frog ganglion cells undergo cell death by apoptosis afteraxotomy.  A:  Light micrograph of a portion of the nasal region of aNissl-stained retina of the frog   Rana pipiens , showing closely packedsomata in the ganglion cell layer (GCL). Large ganglion cells areindicated by arrows. B:  A similar region in the nasal retina 6 monthsafter cutting the optic nerve, showing a greatly reduced cell densityand changes in shape of the somata. Many cell bodies are now bipolar(arrows). C: Higher magnification of a retina at 8 weeks after axotomyshowing apoptotic bodies (arrows).  D:  Transverse section of frog ret-ina stained with ApopTag kit, showing labeled nuclei (arrowheads)and apoptotic bodies (arrows) in the GCL, 6 weeks after axotomy. Theinner plexiform layer (IPL) and inner nuclear layer (INL) are alsoindicated. Scale bar  50  m in A,B,D; 15  m in C. 649bFGF RESCUES FROG RETINAL GANGLION CELLS  icant effect on cell survival at 12 weeks. Control injectionsof PBS also had no effect (Fig. 4B).  Application of bFGF rescues ganglion cells labeledwith TDA.  Because the Nissl stain labels all cell bodiesin the GCL, the cell counts described above include somedisplaced amacrine cells (estimated at 8–11 cells per 0.01mm 2 ). However, RGCs can be identified unequivocallyusing retrograde labeling with Texas red dextran amine(TDA; Fig. 5). This technique consistently labels 60–75%of the cells in the GCL and confirms that large numbers of RGCs are lost at 12 weeks after axotomy (Fig. 5B) andthat two applications of bFGF to the nerve stump rescuemany of these cells (Fig. 5C). Quantitative counts made intwo different regions of the retina showed statisticallysignificant 40–50% increases in numbers of labeled RGCs(Fig. 6A,B), increases that are similar to those observedusing Nissl staining. Cells of Dogiel in the inner nuclearlayer were probably included in the TDA-labeled countsbut not in the Nissl-stained counts. Although no attemptwas made to determine whether the numbers of Dogielcells are differentially affected by axotomy or bFGF treat-ment, they represent only 1–5% of the normal RGC pop-ulation (Frank and Hollyfield, 1987). Morphological changes in GCL neuronsfollowing axotomy and bFGF treatment We were interested in determining whether cells in theGCL were altered in size or shape by axotomy and bFGFtreatment. Two parameters of GCL neurons were mea-sured, the equivalent circle diameter (ECD; diameter of acircle with an area equal to that of the cell body) and theform factor (the longest dimension of the cell body dividedby its width). At 12 weeks after axotomy, there weresignificant changes in the histograms of both parameters,averaged over four retinas. There was a significant de- Fig. 2. Diagram of frog retina, showing contours of cell density inthe ganglion cell layer (number of cells/0.01 mm 2 ; data obtained fromHumphrey, 1987). Eight areas were chosen for cell counts (largesquares); three 0.01 mm 2 samples were taken from each area (smallsquares).Fig. 3. Differential rates of cell death in the frog retina afteraxotomy. Graphs of the mean number of cells in 0.01 mm 2 samples of different regions of central and peripheral retina, plotted against timeafter axotomy. Error bars indicate the standard error of the mean.Four preparations were counted for each data point, with the excep-tion of controls and 26 week preparations, where n  5. Preparationsin which the optic nerve was cut are indicated by triangles, those inwhich it was only crushed are indicated by circles. 650 R.E. BLANCO ET AL.
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks