Vision Research 39 (1999) 685–697

Colchicine causes excessive ocular growth and myopia in chicks

Andy J. Fischer a,*, Ian G. Morgan b, William K. Stell a

a Department of Anatomy and Lions’ Sight Centre, The Uni6ersity of Calgary, Faculty of Medicine, 3330 Hospital Dr. NW, Calgary,Alberta T2N 4N1, Canada

b Centre for Visual Sciences, Research School of Biological Sciences, Australian National Uni6ersity, GPO Box 475, Canberra ACT 2601, Australia

Received 16 January 1998; received in revised form 23 May 1998

Abstract

Colchicine has been reported to destroy ganglion cells (GCs) in the retina of hatchling chicks. We tested whether colchicineinfluences normal ocular growth and form-deprivation myopia, and whether it affects cells other than GCs. Colchicine greatlyincreased axial length, equatorial diameter, eye weight, and myopic refractive error, while reducing corneal curvature. Colchicinecaused DNA fragmentation in many GCs and some amacrine cells and photoreceptors, ultimately leading to the destruction ofmost GCs and particular sub-sets of amacrine cells. Colchicine-induced ocular growth may result from the destruction of amacrinecells that normally suppress ocular growth, and corneal flattening may result from the destruction of GCs whose central pathwaynormally plays a role in shaping the cornea. © 1998 Elsevier Science Ltd. All rights reserved.

Keywords: Form deprivation myopia; Ocular growth; Amacrine cell; Ganglion cell; Chicken

1. Introduction

The shape and size of the eye are crucial to thequality of vision. Emmetropia is achieved when thedistance between the retina and the lens is matched tothe combined refractive power of the lens and cornea sothat distant images are focussed upon the retina withaccommodation relaxed. If the focal plane of an imagefalls in front of the retina the eye is myopic (near-sighted), whereas if the focal plane is behind the retinathe eye is hyperopic (far-sighted). Early in life, growthof the eye is regulated precisely so that emmetropia isachieved. The process of emmetropization can be per-turbed by depriving the eye of patterned images, whichresults in elongation of the vitreous chamber and my-opia (Wallman, 1993). The paradigm of form-depriva-tion myopia (FDM), commonly studied in chicks, hasprovided ample evidence that visual experience is re-quired to properly guide ocular growth and that thisregulation of growth is mediated by retinal activity(Wallman, Gottlieb, Rajaram & Fugate-Wentzek, 1987;Troilo, Gottlieb & Wallman, 1987).

One question that has been addressed by researchersis whether accommodation and retinal connections tothe brain are required for emmetropization. Disablingof accommodative mechanisms and interruption of reti-nal connections to the brain by optic nerve section,intravitreal application of tetrodotoxin, lesions to thenucleus of Edinger-Westphal, removal of the ciliaryganglion, or severing of the ciliary nerves have little orno effect on emmetropization or FDM (Wallman,Rosenthal, Adams & Romagnona, 1981; Troilo, Got-tlieb & Wallman, 1987; Schaeffel, Glasser & Howland,1988; Schaeffel, Troilo, Wallman & Howland, 1990;Troilo, 1990; Lin & Stone, 1991; Troilo & Wallman,1991; Wildsoet & Howland, 1991; McBrien, Moghad-dam, Cottriall, Leech & Cornell, 1995). These findingssuggest that the mechanisms of visual guidance ofocular growth are intrinsic to the eye and do notrequire active accomodation or communication withhigher visual centers.

The purpose of this study was to explore anothermodel for removal of retinal input to the brain, andthereby to test independently the hypothesis that retinalconnections to the brain are required to modulateocular growth. Colchicine has been reported to destroymost retinal ganglion cells, while sparing intrinsic reti-nal neurons (Morgan, 1981). Therefore, we injected

* Corresponding author. Tel.: +1 403 2208724; fax: +1 4032832700; e-mail: ajfische@acs.ucalgary.ca

0042-6989/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.

PII: S0042-6989(98)00178-3

A.J. Fischer et al. / Vision Research 39 (1999) 685–697686

colchicine into the eyes of newly hatched chicks, mea-sured their refractive state and ocular dimensions 2weeks after treatment, tested whether they could re-spond to form-deprivation, and characterized damagecaused by colchicine using a variety of cytochemicalmarkers. Surprisingly, we found that colchicine causedlarge increases in the size of eyes, comparable to thosecaused by form-deprivation, accompanied by cornealflattening. In addition, colchicine not only destroyedganglion cells but also amacrine cells, including somethat have been implicated in the visual regulation ofocular growth. We conclude that colchicine not onlyaffects retinal ganglion cells, but also damages or de-stroys specific subsets of retinal amacrine cells. It islikely that some of the lesioned amacrine cells arenormally responsible for the visual control of vitrealchamber elongation, whereas the ablated ganglion cellscontrol corneal shape via central and peripheral neuralpathways.

2. Material and methods

2.1. Animals

Newly hatched male leghorn chickens (Gallus gallusdomesticus) were obtained from Lillydale Hatchery(Linden, Alberta) and kept on a cycle of 12 h light, 12h dark (lights on at 07:00 h). Illumination was pro-vided by 100 W incandescent light bulbs, resulting inirradiance levels 50.8 cpd/m2, depending on the direc-tion of gaze. Prior to experimentation, chicks were heldfor 1 week in a stainless steel brooder, at about 25oC.Chicks received water and Purina™ chick starter adlibitum.

2.2. Intraocular injections

Chicks were anesthetized with 1.5% halothane in50% N2O and 50% O2 prior to injection. Injectionswere made into the vitreous chamber using a 25-mlHamilton syringe with a 26 gauge needle. Penetrationof the needle was consistently made into the dorsalquadrant of the eye. Shortly after the chicks werehatched, the right eye (control) was injected with 20 mlof vehicle (sterile saline) and the left eye (treated) wasinjected with 20 ml of sterile saline or 20 ml of 25 mM(0.5 nmol or 0.2 mg) colchicine (Research BiochemicalsInternationl, Natick, MA). Assuming that the volumeof liquid-vitreous within an eye was 150 ml, the initialmaximum concentration of colchicine presented to theretina was about 3 mM. One, 3 and 7 days later,pupillary responses were tested by shining a flashlightover the dark-adapted eye of the chick. Seven daysafter treatment, eyes were either covered with atranslucent goggle or left open. At 7 days (i.e. before

form-deprivation) and 14 days of age (i.e. after form-deprivation), eyes were refracted using a streakretinoscope and trial lenses without cycloplegia oranesthesia. At 14 days of age chicks were killed bychloroform inhalation, eyes removed from the orbit,and most of the attached connective tissues and mus-cles trimmed away. Eyes were weighed (to the nearestmilligram) using a Mettler PL200 digital balance, andthe equatorial diameter (at the most narrow part of theeye) and axial length (at the apex of the cornea to thebase of the posterior eye) were measured using digitalcalipers.

2.3. Fixation and sectioning

Enucleated eyes were hemisected equatorially andthe gel vitreous removed from the posterior eye cup.Eye cups were fixed for 30 min at 20oC in 4%paraformaldehyde plus 3% sucrose in 0.1 M phosphatebuffer, pH 7.4. Some samples were fixed for 24 h at4oC, since these conditions were better for immunola-belling with antisera to GAD-65, GABA or serotonin.Fixed samples were washed three times in PBS (phos-phate-buffered saline; 0.05 M phosphate buffer, 195mM NaCl, pH 7.4), cryoprotected in PBS plus 30%sucrose, soaked in embedding medium (OCT-com-pound; Tissue-Tek) for 10 min, and freeze-mountedonto aluminum sectioning blocks. Vertical sectionsnominally 12–14 mm thick were cut consistently fromthe posterior pole of the eye in the nasotemporal plane,near the dorsal portion of the pecten, and thaw-mounted onto gelatin-coated glass slides. Sections fromcontrol and treated eyes from the same individual wereplaced consecutively on each slide to ensure equalexposures to reagents. Sections were air-dried, ringedwith rubber cement, and stored at −20oC until use.

2.4. Immunocytochemistry

Sections were washed three times in PBS, coveredwith primary antibody solution (Table 1; 150 ml ofantiserum diluted in PBS plus 5% normal goat serum,0.3% Triton X-100, and 0.01% NaN3), and incubatedfor about 24 h at 20oC in a humidified chamber. Theslides were washed three times in PBS, covered withsecondary antibody solution (150 ml of 1:1500 Cy3-conjugated goat-anti-rabbit IgG or mouse IgG, Amer-sham), and incubated for at least 2 h at 20oC in ahumidified chamber. Finally, samples were washedthree times in PBS, rubber cement removed from theslides, and coverslips mounted on 4:1 (v/v) glycerol towater for observation under an epifluorescence micro-scope using a rhodamine filter combination. Photo-graphs were taken on T-Max 400 film (Kodak) andnegatives developed in T-Max Developer (Kodak) asper the manufacturer’s instructions.

A.J. Fischer et al. / Vision Research 39 (1999) 685–697 687

Table 1Antisera

Source Working dilutionAntiserum Antigen Species/type

Dr J.H. Rogers 1:1000pURXR.3 Calretinin Rabbit polyclonalRabbit polyclonal Dr M. Epstein1465 Choline acetyltransferase (ChAT) 1:1000

Dr J.-M. FritschyMouse monoclonal 1:50GABAA receptorBD-171:1000Dr J. Walsh8305034 Glucagon Rabbit polyclonal

Rabbit polyclonal Dr C. Brandon634 Glutamic acid decarboxylase (GAD-65) 1:1000Dr M. ChiquetMouse monoclonal 1:70Bipolar cell markerM3Dr J. Walsh 1:10001473 Met-enkephalin (ENK) Rabbit polyclonal

1:50Hybridoma BankMouse monoclonalNeurofilament-associated antigen3A10Mouse monoclonal SigmaP-3171 Parvalbumin 1:1600Mouse monoclonal AmershamRPN536 Protein kinase C (PKC) a and b isoforms 1:50

1:800Inc StarRabbit polyclonalSerotonin (5-HT)5-HTTransforming growth factor-b2 (TGF-b2) Rabbit polyclonal Santa Cruz 1:1000sc-90

1:1000Dr W. Tankc16 Tyrosine hydroxylase (TH) Rabbit polyclonal1:80Vasoactive intestinal polypeptide (VIP) Rat monoclonalVP31 Dr A. Buchan

2.5. Histology

Slides were warmed to 20oC, washed three times inPBS, and incubated under 200 ml 0.1% (w/v) toluidineblue plus 0.2% (w/v) Na2B4O7 in dH2O at pH 11.4 forabout 2 min. The stain was drained away, slides washedthree times in PBS, and mounted, as described above,for microscopy in transmitted white light.

2.6. Labelling of fragmented DNA

Retinal sections were obtained as described abovefrom chicks 1, 3, 5 and 7 days after colchicine-treat-ment. Slides were warmed to 20oC and washed once inPBS, followed by one wash in PBS plus 0.3% TritonX-100, and two more washes in normal PBS. Sectionswere then covered with 150 ml of incubation medium(0.5 nmol Cy3-conjugated dCTP, 20 units of 3%-terminaldeoxynucleotidyl transferase (Amersham), 100 mMsodium cacodylate, 2 mM CoCl2, and 0.25 mM b-mer-captoethanol, in distilled water at pH 7.2) and incu-bated for 1 h in a humidified chamber at 37oC. Sectionswere then washed three times in PBS, mounted in 4:1(v/v) glycerol to water, and coverslips added for obser-vation by epifluorescence using a rhodamine filtercombination.

2.7. Measurements of corneal cur6ature

Photographs were made of enucleated eyes viewedfrom the side. The negatives of these photographs weredigitized to a resolution of 20 pixels per mm. UsingAdobe Photoshop 4.0™, a rectangular area wasblocked out, with one corner originating at the apex ofthe cornea while the opposing corner was set at thebase of the cornea. The number of pixels in the heightand width of the rectangle was then determined, and aheight to width ratio calculated as a function represen-tative of corneal curvature.

2.8. Measurements, cell counts, and statistical analyses

Errors were calculated as the standard deviation ofeach sample which was comprised of at least six indi-viduals per group. To assess the differences betweendata from treated and control eyes, we used a pairedtwo-tailed Student t-test. To assess the differences be-tween eyes from multiple different treatments, we useda two-way ANOVA (In Stat© for MacIntosh). Percent-age IPL depth was calculated as the distance from theborder between the inner plexiform layer (IPL) andinner nuclear layer (INL), divided by the total IPLthickness, multiplied by 100. All thickness measure-ments were made from photomicrographs of centralretina, while all cell counts of at least 50 cells perindividual on at least four different sections were madeunder the microscope.

3. Results

3.1. Pupillary responses

One day after treatment with colchicine, all chicks(n=32) had pupillary responses, although the re-sponses in treated eyes were slightly reduced. Three and7 days after treatment, eyes injected with saline hadstrong pupillary constriction when exposed to brightlight, while there was no response in colchicine-treatedeyes. Furthermore, by 3 days after treatment chickswere blind in their treated eye, as they did not respondto an approaching hand. All chicks were still able tofeed normally.

3.2. Changes in gross retinal morphology

Histological and immunocytochemical examinationswere made only 14 days after treatment, when the

A.J. Fischer et al. / Vision Research 39 (1999) 685–697688

Fig. 1. Vertical sections of retina that have been stained with toluidine blue from (a) control and (b) colchicine-treated eyes. Abbreviations: IPL,inner plexiform layer; INL, inner nuclear layer; GCL, ganglion cell layer. Scale bar=50 mm.

colchicine-induced lesion should have been stabilized(Morgan, 1981). Colchicine caused significant losses ofthickness from several retinal layers, particularly theoptic fibre layer (OFL) (28.796.2% of the control;mean9S.D.; n=8) and IPL (55.598.3% of the con-trol). The thickness of the INL was also slightly re-duced to 83.198.6% of the control. Colchicine causedan obvious loss of cells and thickness from the ganglioncell layer (GCL; Fig. 1); the majority of the cellsremaining were assumed to be displaced amacrine cells(Morgan, 1981). The photoreceptor layer and retinalpigmented epithelium (RPE) appeared dysplastic, thinor discontinuous in some parts, while appearing thickerin other regions (Fig. 1).

3.3. Effects of colchicine upon immunohistologicallydistinct populations of retinal neurons

Colchicine caused no detectable loss in labelling in-tensity or distribution of amacrine cells immunoreactivefor GAD-65, GABAA receptor, serotonin, parvalbu-min, or PKC, except for the compression of neuriteswithin the thinned IPL. Colchicine had no apparenteffect upon bipolar cells labelled with monoclonal anti-body M3 or antiserum to PKC, or upon nerve fibers inthe choroid that were immunoreactive for NAA,ChAT, TH or VIP (results not shown).

3.4. Enkephalin

Antiserum to enkephalin (ENK) labelled many cellshaving somata near the middle of the INL, and den-drites in several distinct sublaminae of the IPL (Fig.2a). These cells are the enkephalin-neurotensin-somato-statin-like immunoreactive (ENSLI) amacrine cellswhich have been described previously in avian retina(Brecha, Karten & Laverack, 1979; Watt, Li & Lam,1985; Morgan, Wellard & Boelen, 1994; Watt & Flo-rack, 1994). After exposure to colchicine there was a

substantial loss of ENK-immunoreactive amacrine cells(Fig. 2a), nearly half of which (45.1913.8%; n=6)perished after 14 days. The cells that survived appearednormal, lacking noticeable swelling of their somata ordegeneration of terminal arbors (Fig. 2a).

3.5. Choline acetyltransferase

Antiserum to choline acetyltransferase (ChAT) la-belled many orthotopic and displaced amacrine cells, aswell as two prominent strata in the IPL (Fig. 2b),exactly as described previously (Millar, Ishimoto,Chubb, Epstein, Johnson & Morgan, 1987). Subtypesof cholinergic amacrine cells included: type-I cells, withsomata in the proximal INL and neurites in sublamina2 of the IPL; type-II cells, with somata in the GCL andneurites in sublamina 4 of the IPL; and type-III cells,with somata near the middle of the INL and neuritesdiffusely distributed in sublaminae 1 and 3–5 of theIPL. As reported elsewhere, about half of the type-IIIcells are ENSLI cells (Fischer, Poon, Seltner & Stell,1998). Colchicine had no apparent effect upon type-Iand type-II cholinergic amacrine cells, while inducingsome loss of type-III cells (Fig. 2b). This was foreshad-owed by the results of ENK-immunolabelling, in whichmany type-III/ENSLI cells were destroyed bycolchicine.

3.6. Tyrosine hydroxylase

Antiserum to tyrosine hydroxylase (TH) labelledsparsely distributed cell bodies with neurites in threedifferent strata in the IPL, at 0–10, 35, and 75% depth(Fig. 2c), exactly as described previously (Su & Watt,1987). Only about 10% (10.996.5%; n=6) of TH-im-munoreactive amacrine cells survived exposure to col-chicine, and consequently there were massive losses ofTH-positive neurites from the IPL; this loss of TH-im-munoreactive cells was uniform across the entire retina(Fig. 2c).

A.J. Fischer et al. / Vision Research 39 (1999) 685–697 689

Fig. 2. Vertical sections of retinas that have been treated with saline (left panel) or colchicine (right panel) and labelled with antibodies directedagainst (a) enkephalin; (b) choline acetyltransferase; (c) tyrosine hydoxylase; (d) vasoactive intestinal polypeptide; (e) glucagon; (f) calretinin; (g)neurofilament-associated antigen; and (h) transforming growth factor-b2. Large arrow heads indicate the inner limiting membrane and smallarrow heads indicate the proximal and distal borders of the IPL. Abbreviations: IPL, inner plexiform layer; INL, inner nuclear layer; GCL,ganglion cell layer. Scale bar=50 mm.

A.J. Fischer et al. / Vision Research 39 (1999) 685–697690

Fig. 2. (Continued)

3.7. Vasoacti6e intestinal polypeptide

Antiserum to vasoactive intestinal polypeptide (VIP)labelled puncta in the perinuclear cytoplasm of manycell bodies near the border of the INL and IPL, in

addition to sparsely distributed neurites at 0–10, 35,and 70% IPL depth, as described previously (Fig. 2d;Brecha, 1983; Seltner & Stell, 1995). Colchicine causeda substantial loss of VIP-containing cell bodies in theINL and neurites in the IPL (Fig. 2d).

A.J. Fischer et al. / Vision Research 39 (1999) 685–697 691

3.8. Glucagon

Antiserum to glucagon labelled amacrine cells havingsomata at the IPL/INL border, one densely innervatedstratum at 0–10% IPL depth, and one sparsely inner-vated stratum at 35% IPL depth (Fig. 2e), as describedpreviously (Kiyama, Katayama-Kumoi, Kimmel, Stein-busch, Powell, Smith & Tohyama, 1985). More than90% (90.991.0%; n=6) of glucagon-immunoreactiveamacrine cells were destroyed by colchicine, leaving veryfew glucagon-containing neurites in the IPL; this loss ofglucagon-immunoreactive cells was uniform across theentire retina (Fig. 2e).

3.9. Calretinin

Antiserum to calretinin labelled most cells in the GCL,amacrine cells, and horizontal cells, as well as a subsetof bipolar cells with somata located towards the middleof the INL (Fig. 2f). Both plexiform layers and the OFLwere filled with dendrites and/or axons that were im-munoreactive for calretinin. Calretinin-immunoreactiv-ity could be distinguished in at least 12 strata within theIPL (Fig. 2f). This distribution of calretinin is consistentwith previous reports (Ellis, Richards & Rogers, 1991;Rogers, Khan & Ellis, 1989; Rogers, 1989). Two weeksafter exposure to colchicine, little or no calretinin im-munoreactivity remained in the GCL and OFL (Fig. 2f).This suggests that the type-II cholinergic cells displacedto the GCL, which were unaffected by colchicine, are notimmunoreactive for calretinin. Colchicine had little orno effect upon calretinin-immunoreactive horizontal,bipolar, and amacrine cells. Furthermore, there ap-peared to be little change in the distribution of calretinin-immunoreactivity in the IPL and OPL, except for thecompression of strata in the thinned IPL.

3.10. Neurofilament-associated antigen (NAA)

Antiserum to neurofilament-associated antigen(NAA) robustly labelled the axons of ganglion cells inthe OFL and of displaced ganglion cells or efferenttarget cells at the INL/IPL border (Fig. 2g). In controlretinas, all ganglion cells were only weakly immunoreac-tive for NAA (results not shown). Colchicine caused amassive loss of NAA-immunoreactivity from the OFLand significant depletion of NAA-immunoreactive fibresat the INL/IPL border (Fig. 2g). The few ganglion cellsthat survived exposure to colchicine became hyperim-munoreactive for NAA and often appeared swollen andabnormal (Fig. 2g).

3.11. Transforming growth factor-b2 (TGF-b2)

Antiserum to transforming growth factor-b2 (TGF-b2) labelled sparsely distributed amacrine cells, with

somata located near the border of the IPL and INL, andmost cells in the GCL (Fig. 2h), as observed previously(Gudgeon, Fischer & Stell, 1998). Colchicine had noeffect upon TGF-b2-immunoreactive amacrine cells, butabolished most TGF-b2-immunoreactivity in the GCL(Fig. 2h). Surviving ganglion cells became hyperim-munoreactive for TGF-b2 and often appeared swollenand abnormal (Fig. 2h).

3.12. Colchicine-induced DNA fragmentation

Twenty-four hours after exposure to colchicine therewas no detectable DNA fragmentation in treated retinas.

Fig. 3. Vertical sections of retinas treated with colchicine and labelledfor DNA fragmentation (a) before, and (b) 3 and (c) 5 days aftertreatment. Abbreviations: IPL, inner plexiform layer; INL, innernuclear layer; GCL, ganglion cell layer; ONL, outer nuclear layer.Scale bar=50 mm.

A.J. Fischer et al. / Vision Research 39 (1999) 685–697692

However, 3 days after treatment the nuclei of manyamacrine cells were labelled for fragmented DNA,while only a few nuclei of photoreceptors and ganglioncells were labelled (Fig. 3b). Five days after treatment,fragmented DNA was detected in the nuclei of manyganglion cells and a few amacrine cells, but was nolonger present in the nuclei of photoreceptors (Fig. 3c).Seven days after treatment, fragmented DNA was nolonger detected in any retinal cells (results not shown).

3.13. Ocular growth in colchicine-treated eyes

Fourteen days after treatment with colchicine, eyeshad grown excessively and become myopic. In compari-son to control eyes, the axial length, equatorial diame-ter, weight, and negative refractive error weresignificantly (PB0.0001) greater in treated eyes (Fig.4). On average, colchicine-treated eyes were 10.292.9%longer, 5.091.9% wider, and weighed 20.694.9%more than contralateral control eyes. These increases ineye size were not significantly different from the in-creases caused by 7 days of form-deprivation of saline-injected control eyes (Fig. 4). However, the amount ofmyopia that resulted from colchicine treatment wassignificantly (P=0.0006) less than that caused by form-deprivation. Furthermore, form-deprivation did notfurther increase eye size or refractive error caused bycolchicine (Fig. 4). Colchicine also caused significant(P=0.0001; n=7) flattening of the cornea (Fig. 5). Theheight to width ratio of corneas from saline-injectedeyes was 0.64590.056 (n=7), while that of corneasfrom colchicine-treated eyes was 0.49490.044 (n=7).The height of treated corneas was 39.098.3% less thanthat of controls.

4. Discussion

4.1. Colchicine-induced ocular enlargement

Colchicine-treated eyes received normal visual stim-uli, which are required for normal growth of the eye,but were unable to emmetropize. Therefore, among thedamage elicited by colchicine was the disruption ofpathways that promote emmetropization. Colchicinealso resulted in the activation of growth-promotingpathways or the destruction of growth-suppressingpathways, thereby allowing ocular growth to rununchecked or by default. These changes in oculargrowth may have been caused by colchicine-induceddamage to the RPE, photoreceptors, ganglion cells,and/or subsets of amacrine cells, or by undetecteddamage to ocular tissues.

Colchicine-treated retinas were unable to send thebrain normal information regarding the visual environ-ment, because of the destruction of most ganglion cells.

It is unlikely that this loss of ganglion cells causedexcessive growth or prevented emmetropization, as op-tic nerve sectioning also causes destruction of mostganglion cells but has little effect upon normal oculargrowth or FDM (McBrien, Moghaddam, Cottriall,Leech & Cornell, 1995; Troilo, Gottlieb & Wallman,1987; Wildsoet & Pettigrew, 1988). In fact, lesioning ofthe optic nerve results in smaller, hyperopic eyes(Troilo, Gottlieb & Wallman, 1987), although it is notclear whether this is caused by loss of ganglion cellinput to the brain or damage to intrinsic retinal circuitsdue to surgery.

Colchicine also damaged or destroyed some photore-ceptors. It is unlikely that photoreceptors participatedirectly in the visual modulation of ocular growth,although damage to photoreceptors that activate pref-erentially a critical subset of retinal pathways might inprinciple disrupt emmetropization. Furthermore, thedestruction of many photoreceptors with tunicamycindoes not affect normal eye growth or the progression ofFDM (Ehrlich, Sattayasai, Zappia & Barrington, 1990).The destruction of photoreceptors and the RPE withformoguanamine also does not affect normal eyegrowth, but does prevent FDM (Oishi & Lauber, 1988).Therefore, it is unlikely that colchicine-induced damageto photoreceptors caused excessive ocular growth or theinability of treated eyes to emmetropize.

The accelerated ocular growth caused by colchicinemay have resulted from the destruction of retinalamacrine cells. It is possible that the normal suppressivecontrol of ocular growth requires the activity ofamacrine cells that are affected by colchicine, includingthose that contain VIP, enkephalin and ChAT, TH orglucagon, or other amacrine cells for which we did notprobe. Pharmacological studies have implicated severalsubsets of amacrine cells in the progression of FDM,including those that produce dopamine (Stone, Lin,Laties & Iuvone, 1989), VIP (Stone, Laties, Raviola &Weisel, 1988; Seltner & Stell, 1995), ENK (Seltner,Rohrer, Grant & Stell, 1997), and acetylcholine (Stone,Lin & Laties, 1991; McBrien, Moghaddam & Reeder,1993; Leech, Cottriall & McBrien, 1995). These studieshave suggested that retinal cells that release VIP, ENKor acetylcholine promote ocular growth during form-deprivation, while cells that release dopamine suppressgrowth. However, in the quisqualate-treated retina,cells containing VIP, ENK or acetylcholine are com-pletely destroyed, yet these eyes remain emmetropic andare capable of responding to form-deprivation, therebysuggesting that these cells do not contribute to visuallyguided ocular growth (Fischer, Poon, Seltner & Stell,1998; Fischer, Miethke, Morgan & Stell, 1998). Incontrast, glucagon and TH-immunoreactive amacrinecells are relatively unaffected by quisqualate (Fischer,Poon, Seltner & Stell, 1998), but are eliminated fromcolchicine-treated retinas (present study). There is no

A.J. Fischer et al. / Vision Research 39 (1999) 685–697 693

Fig. 4. Colchicine-induced changes in (a) axial length; (b) equatorial diameter; (c) refractive error; and (d) eye weight. Significance (*PB0.0001)of difference between means of treated samples and controls was assessed by using a one-way Student t-test. Abbreviation: FD, form-deprived.

evidence that glucagon-immunoreactive amacrine cellsmodulate ocular growth. The evidence for a role ofdopamine and TH-immunoreactive cells in growth-reg-ulation and FDM is mixed, but the concurrent destruc-

tion of TH cells and induction of ocular enlargement bycolchicine is consistent with other evidence favoringsuch a role (Stone, Lin, Laties & Iuvone, 1989; Rohrer,Spira & Stell, 1993).

A.J. Fischer et al. / Vision Research 39 (1999) 685–697694

Fig. 5. Photographs of eyes that have been treated with (a & c) saline; (b) form-deprived; or (d) treated with colchicine. Scale bar=5 mm.

Alternatively, colchicine may have destroyed or influ-enced the activity of cells in the choroid or sclera thatnormally suppress ocular growth. However, it is coun-ter-intuitive to expect that colchicine-induced destruc-tion or suppression of normal cellular activities, such asmitosis or secretion of synthesized products, could re-sult in excessive ocular growth. It also seems unlikelythat a single dose of colchicine could elicit an effectover the 2-week observation period used in our experi-ments unless there was a permanent removal of cellswithout any replacement.

Form-deprivation did not further enhance the nega-tive refractive error or the increased size of colchicine-treated eyes. It is possible that the mechanisms thatpromote ocular growth were ‘saturated’ in both form-deprived and colchicine-treated eyes, so that maximumrates of growth occurred in both conditions. If maxi-mum rates of growth were achieved in both form-de-prived and colchicine-treated eyes, then eyes shouldhave been equally large for both treatments, and form-deprivation of colchicine-treated eyes should not haveresulted in further enhancement of eye size. Indeed, wefound that the size of eyes that were form-deprived,colchicine-treated, and colchicine-treated and form-de-prived were nearly equivalent. This suggests that col-chicine and form-deprivation activate the same

mechanism that enhances ocular growth. Alternatively,growth-promoting mechanisms may not have been sat-urated, and colchicine may have destroyed the retinalcircuitry responsible for form-deprivation-inducedgrowth, thereby rendering the treated eyes unresponsiveto form-deprivation.

4.2. Colchicine-induced corneal flattening

Colchicine caused a pronounced flattening of thecornea. This accounted at least in part for the smallerdegree of myopia in colchicine-treated open eyes thanin saline-treated form-deprived eyes, as the ocular di-mensions of these eyes were not significantly differentand in themselves should have produced similar refrac-tive errors. Flattening of the cornea would result inreduced refractive power, thereby decreasing the refrac-tive error of these enlarged eyes. It is also possible thatchoroidal thickening (Wallman, Wildsoet, Xu, Gottlieb,Nickla, Marran, Krebs & Christensen, 1995; Wildsoet& Wallman 1995) could have decreased refractive errorof colchicine-treated eyes, but this possibility was nottested in these experiments.

Corneal flattening may have resulted from the col-chicine-induced loss of ganglion cells. Flattening of thecornea was observed by Troilo & Wallman (1991) after

A.J. Fischer et al. / Vision Research 39 (1999) 685–697 695

optic nerve sectioning, and by McBrien, Moghaddam,Cottriall, Leech & Cornell (1995) after chronic ad-ministration of tetrodotoxin. Although colchicine,tetrodotoxin, and optic nerve lesioning all elicit differ-ent effects and may perturb different functions withinthe eye, they share suppression or abolition of gan-glion cell activity and transmission to the brain.Therefore, we propose that corneal growth is regu-lated by activity in one or more classes of retinalganglion cell, presumably through some visual cen-ter(s) in the brain and outflow via the cilary ganglionand nerves.

4.3. Colchicine-induced destruction of retinal cells

The mechanisms underlying the destruction of reti-nal cells by colchicine remain uncertain. FragmentedDNA was detected in both amacrine and ganglioncells, suggesting that these cells became apoptotic af-ter exposure to colchicine. Colchicine causes the dis-assembly of microtubules, and thereby interferes withmany cellular functions. One function disrupted bycolchicine is axonal transport (Hanson & Edstrom,1978), and ganglion cells might be expected to perishbecause both anterograde and retrograde transport(e.g. of trophic factors) were disabled. The reductionin effect of colchicine in the eyes of older chicks(Morgan, 1981), might reflect a reduced trophic de-pendency. However, in the chick, natural death ofretinal ganglion cells and establishment of retinotectalconnections occur before hatching (Rager, 1976;Rager & Rager, 1978; Rager & Oeynehausen, 1979).Competition for trophic targets would, therefore, becompleted long before we observed colchicine-inducedcell death. This does not, however, exclude the possi-bility that these cells have a continued trophic depen-dency after tectal connections have been established.This in fact is likely, since axotomy usually results inthe death of ganglion cells (Bray, Villegas-Perez, Vi-dal-Sanz, Carter & Aguayo, 1991).

Colchicine probably prevented the anterogradetransport of proteins away from the cell bodies ofresidual ganglion cells. The accumulation of NAAand TGF-b2-immunoreactivity in some ganglion cellsafter exposure to colchicine suggests that protein syn-thesis was unaffected while the transport of synthe-sized products away from the cell body was disabled.Residual ganglion cells are likely to be non-functionalor dysfunctional, as axonal transportation was pre-vented. The consequent accumulation of TGF-b2-im-munoreactivity in ganglion cell bodies implies thatTGF-b2 is normally exported out of the retina alongaxons, or into the IPL in dendrites. It has been ar-gued that the blockade of anterograde transport andsubsequent build-up of proteins in cell bodies resultsin apoptosis and that this is somehow related to the

degree of axon myelination and requirements for theexchange of lipid materials between axons and oligo-dendrocytes (Droz, 1979; Morgan, 1981).

It remains uncertain why only certain sub-popula-tions of amacrine cells perished after exposure to col-chicine. There is no precedent for dependence ofintrinsic retinal neurons on cytoplasmic transport oftrophic factors. However, there is also no evidenceagainst such dependence and thus we cannot excludethe possiblity that colchicine caused the death ofamacrine cells by depriving them of retrogradedlytransported trophic factors. All types of amacrinecells that were affected by colchicine, in particular thestrongly affected TH and glucagon-immunoreactivecells, had broad dendritic arbors. It is possible thatthe wide dendritic arbors and consequently increasedrequirements for anterograde and retrograde transportrendered these amacrine cells more susceptible to col-chicine-induced cell death. For example, colchicine-in-duced cytoskeletal alterations in rat cerebellar granulecells have been postulated to directly initiate apopto-sis (Bonofoco, Ceccatelli, Manzo & Nicotera, 1995).

4.4. Summary and conclusions

Colchicine destroyed most ganglion cells, as well assubtypes of amacrine cells including those thought toparticipate in visual guidance of ocular growth. Col-chicine also damaged the RPE and photoreceptors.As a result of the damage caused by colchicine, eyeswere unable to emmetropize, grew excessively, becamemyopic, and had flattened corneas. We propose thatcolchicine-induced elongation of the vitreous chamberresults from the destruction of one or more subsets ofamacrine cells that normally suppress growth in re-sponse to visual stimuli, and that colchicine-inducedcorneal flattening results from the destruction of oneor more subsets of ganglion cells that normally pro-mote proper corneal shape.

Acknowledgements

This work was supported by the Silver AnniversaryGraduate Fellowship from the University of Calgary,and by studentships from the Pharmaceutical Manu-facturers Association of Canada-Medical ResearchCouncil and The Alberta Heritage Foundation forMedical Research to Andy Fischer; by grants fromThe Medical Research Council of Canada, the RoyAllen Fund, the Marigold Foundation, and the Ed-win L. and John E. Gustus Endowment in VisionDisorders at the University of Calgary to W.K. Stell;and by a Scientific Visits Grant from the AustralianAcademy of Science to I.G. Morgan.

A.J. Fischer et al. / Vision Research 39 (1999) 685–697696

References

Bonofoco, E., Ceccatelli, S., Manzo, L., & Nicotera, P. (1995).Colchicine induces apoptosis in cerebellar granule cells. Experi-mental Cell Research, 218, 189–200.

Bray, G. M., Villegas-Perez, M. P., Vidal-Sanz, M., Carter, D. A., &Aguayo, A. J. (1991). Neuronal and non-neuronal influences onretinal ganglion cell survial, axonal regrowth, and connectivityafter axotomy. Annals of the New York Academy of Sciences, 613,214–228.

Brecha, N. (1983). Retinal neurotransmitters: histochemical and bio-chemical studies. In P. C. Emson, Chemical Neuroanatomy. NewYork: Raven Press, 85–129.

Brecha, N., Karten, H. J., & Laverack, C. (1979). Enkephalin con-taining amacrine cells in the avian retina: immunohistologicallocalization. Proceedings of the National Academy of Science USA,76, 3010–3014.

Droz, B. (1979). How axonal transport contributes to maintenance ofthe myelin sheath. Trends in Neuroscience, 2, 146–148.

Ellis, J. H., Richards, D. E., & Rogers, J. H. (1991). Calretinin andcalbindin in the retina of developing chick. Cell and TissueResearch, 264, 197–208.

Erhlich, D., Sattayasai, J., Zappia, J., & Barrington, M. (1990).Effects of selective neurotoxins on eye growth in the young chick.In E. D. Bock, & K. Widdows, Myopia and the control of oculargrowth, Vol. 155. Chichester: Wiley. Ciba Foundation Sympo-sium, 63–84.

Fischer, A. J., Poon, J., Seltner, R. L. P., & Stell, W. K. (1998).Immunocytochemical characterization of NMDA and quisqualicacid-induced excitotoxicity in the chick retina. The Journal ofComparati6e Neurology, 393, 1–15.

Fischer, A. J., Miethke, P., Morgan, I. G., & Stell, W. K. (1998).Cholinerigic amacrine cells are not required for the progressionand atropine-mediated supression of form-deprivation myopia.Brain Research, 794, 48–60.

Gudgeon, J. H. R., Fischer, A. J., & Stell, W. K. (1998). Localizationof TGF-b2 to the ocular tissues of the chick. (in preparation).

Hanson, M., & Edstrom, A. (1978). Mitosis inhibitors and axonaltransport. International Re6iews of Cytology, 7(Suppl.), 373–402.

Kiyama, H., Katayama-Kumoi, Y., Kimmel, J., Steinbusch, H.,Powell, J. F., Smith, A. D., & Tohyama, M. (1985). Threedimensional analysis of retinal neuropeptides and amines in thechick. Brain Research Bulletin, 15, 155–165.

Leech, E. M., Cottriall, C. L., & McBrien, N. A. (1995). Pirenzepineprevents form-deprivation myopia in a dose dependent manner.Ophthalmology, Physiology and Optmetry, 15, 351–356.

Lin, T., & Stone, R. A. (1991). Autonomic and visual interactions inthe regulation of eye growth and refraction. In6estigati6e Ophthal-mology and Visual Science, 32(Suppl.), 1202.

McBrien, N. A., Moghaddam, H. O., Cottriall, C. L., Leech, E. M.,& Cornell, L. M. (1995). The effects of blockade of retinal cellaction potentials on ocular growth, emmetropization and formdeprivation myopia in young chicks. Vision Research, 35, 1141–1152.

McBrien, N. A., Moghaddam, H. O., & Reeder, A. P. (1993).Atropine rescues experimental myopia and eye enlargement via anon-accomodative mechanism. In6estigati6e Ophthalmology andVisual Science, 34, 205–215.

Millar, T. J., Ishimoto, I., Chubb, I. W., Epstein, M. L., Johnson, C.D., & Morgan, I. G. (1987). Cholinergic amacrine cells of thechicken retina: a light and electron microscope immunocytochem-ical study. The Journal of Neuroscience, 21, 725–743.

Morgan, I. G. (1981). Intraocular colchicine selectively destroysimmature ganglion cells in chicken retina. Neuroscience Letters,24, 255–260.

Morgan, I. G., Wellard, J. W., & Boelen, M. K. (1994). A role for theenkephalin-immunoreactive amacrine cells of the chicken retina inadaptation to light and dark. Neuroscience Letters, 174, 64–66.

Oishi, T., & Lauber, J. K. (1988). Chicks blinded with formogua-namine do not develop lid suture myopia. Current Eye Research,7, 63–73.

Rager, G. (1976). Morphogenesis and physiogenesis of the retinotec-tal connection in the chicken, II. The retinotectal synapses. Pro-ceedings of the Royal Society of London Series B, 192, 331–352.

Rager, G., & Oeynhausen, B. (1979). Ingrowth and ramification ofretinal fibers in the developing optic tectum of the chick embryo.Experimental Brain Research, 35, 213–227.

Rager, G., & Rager, U. (1978). Systems-matching by degeneration, I.A quantitative electronmicroscopic study of generation and de-generation of retinal ganglion cells in the chicken. ExperimentalBrain Research, 33, 65–78.

Rogers, J. H. (1989). Two calcium-binding proteins mark many chicksensory neurons. Neuroscience, 31, 697–709.

Rogers, J. H., Khan, M., & Ellis, J. H. (1989). Calretinin and otherCaBPs in the nervous system. In R. Pochet, D. E. M. Lawson, &C. W. Heizmann, Calcium binding proteins in normal and tran-formed cells. New York: Plenum.

Rohrer, B., Spira, A., & Stell, W. K. (1993). Apomorphine blocksform-deprivation myopia by a dopamine D2-receptor mechanismacting in retina or pigmented epithelium. Visual Neuroscience, 10,447–453.

Schaeffel, F., Glasser, A., & Howland, H. C. (1988). Accomodation,refractive error and eye growth in chicks. Vision Research, 28,639–657.

Schaeffel, F., Troilo, D, Wallman, J., & Howland, H. C. (1990).Developing eyes that lack accomodation grow to compensate forimposed defocus. Visual Neuroscience, 4, 177–183.

Seltner, R. L. P., Rohrer, B., Grant, V., & Stell, W. K. (1997).Endogenous opiates in the chick retina and their role in formdeprivation myopia. Visual Neuroscience, 14, 801–809.

Seltner, R. L. P., & Stell, W. K. (1995). The effect of vasoactiveintestinal peptide on development of form-deprivation myopia inthe chick: a pharmacological and immunocytochemical study.Vision Research, 35, 1265–1270.

Stone, R. A., Laties, A. M., Raviola, E., & Weisel, T. N. (1988).Increase in retinal vasoactive intestinal polypeptide after eyelidfusion in primates. Proceedings of the National Academy of Sci-ence USA, 85, 257–260.

Stone, R. A., Lin, T., & Laties, A. M. (1991). Muscarinic antagonistseffects on experimental myopia. Experimental Eye Research, 52,755–758.

Stone, R. A., Lin, T., Laties, A. M., & Iuvone, P. M. (1989). Retinaldopamine and form-deprivation myopia. Proceedings of the Na-tional Academy of Science USA, 86, 704–706.

Su, Y. Y. T., & Watt, C. B. (1987). Interaction between enkephalinand dopamine in the avian retina. Brain Research, 243, 63–70.

Troilo, D., Gottlieb, M. D., & Wallman, J. (1987). Visual deprivationcauses myopia in chicks with optic nerve section. Current EyeResearch, 6, 993–999.

Troilo, D. (1990). Experimental studies of emmetropization in thechick. In G. R. Bock, & K. Widdows, Myopia and the Control ofEye Growth, Vol. 155. Chichester: Ciba Foundation Symposium,Wiley, 89–102.

Troilo, D., & Wallman, J. (1991). The regulation of eye growth andrefractive state: an experimental study of emmetropization. VisionResearch, 31, 1237–1250.

Wallman, J. (1993). Retinal control of eye growth and refraction.Progress in Retinal Research, 12, 133–153.

Wallman, J., Gottlieb, M. D., Rajaram, V., & Fugate-Wentzek, L. A.(1987). Local retinal regions control local eye growth and myopia.Science, 237, 73–77.

A.J. Fischer et al. / Vision Research 39 (1999) 685–697 697

Wallman, J., Wildsoet, C., Xu, A., Gottlieb, M. D., Nickla, D. L.,Marran, L., Krebs, W., & Christensen, A. M. (1995). Moving theretina: choroidal modulation of refractive state. Vision Research,35, 37–50.

Wallman, J., Rosenthal, D., Adams, J. I., & Romagnona, L. (1981).Role of accommodation and developmental aspects of experimen-tal myopia in chicks. In H. C. Fledelius, P. H. Alsbirk, & E.Goldschimdt, Proceedings of the Third International Myopia Con-ference. The Hague: Dr W. Junk. Documenta OphthalmologicaProceedings Series, 28, 197–206.

Watt, C. B., & Florack, W. J. (1994). A triple-label analysis demon-strating that enkephalin-, somatostatin- and neurotensin-like im-munoreactivities are expressed by a single population of amacrinecells in the chicken retina. Brain Research, 634, 310–316.

Watt, C. B., Li, H. B., & Lam, D. M. K. (1985). The presence ofthree neuroactive peptides in putative glycinergic amacrine cells ofan avian retina. Brain Research, 348, 187–191.

Wildsoet, C. F., & Howland, H. C. (1991). Chromatic aberrationand accommodation: Their role in emmetropization in the chick.In6estigati6e Ophthalmology and Visual Science, 32(Suppl.),1203.

Wildsoet, C. F., & Pettigrew, J. D. (1988). Experimental myopia andanomalous eye growth patterns are unaffected by optic nervesection in chickens: evidence for local control of eye growth.Clinical Vision Sciences, 3, 99–107.

Wildsoet, C. F., & Wallman, J. (1995). Choroidal and scleral mecha-nisms of compensation for spectacle lenses in chicks. VisionResearch, 35, 1175–1194.

.