vascular casts of experimental subretinal neovascularization in...

Vascular Casts of Experimental SubretinalNeovascularization in Monkeys

Hiroshi Ohkuma,* and Stephen J. Ryan

The origin, course, and pattern of experimental subretinal neovascularization (SRN) in monkeys werestudied with scanning electron microscopy (SEM) of Mercox preparations of the choroidal vascularbed. This technique allowed visualization of the entire circulation of the SRN lesion from both theretinal and scleral aspects. The SEM data is comparable to that of fluorescein angiography, butprovides details not detectable by angiography. The afferent arteriole was traced back to the posteriorciliary artery of origin. In the early stages, the efferent vessels appeared to connect with the chorio-capillaris; whereas later the efferent vessels connected directly to the choroidal venous system. Thestudy of such plastic casts enables more accurate assessment of some aspects of the vascular archi-tecture of the SRN frond. Invest Ophthalmol Vis Sci 24:481-490, 1983

Subretinal neovascularization (SRN) is a patho-logic process common to many disciform maculardiseases.' The ultrastructure of the abnormal bloodvessels found in experimental SRN has been de-scribed by us,2 although we found it difficult to eval-uate the branching patterns and the relationship ofthe new vessels to the choroidal vessels because oftheir extension in three dimensions. To resolve thisproblem, we used a technique of plastic casting of thechoroidal vasculature that permitted clear demon-stration of the ocular vessels.3"7 This technique wasfound to be particularly helpful in the evaluation ofexperimental SRN.8 In the present study, the branch-ing pattern and distribution of neovascular vesselsduring lesion growth and regression were studied byscanning electron microscopy (SEM) of the vascularcasts.

Materials and Methods

Ten rhesus monkeys of both sexes were used inthis study. The method used to produce SRN was thelaser photocoagulation technique previously de-scribed.910 All the animals were examined periodi-

From the Department of Ophthalmology, University of South-ern California School of Medicine, and the Estelle Doheny EyeFoundation, Los Angeles, California.

* Current address: Department of Ophthalmology, Kansai Med-ical University Fumizonocho Moriguchi, Osaka 570, Japan.

Supported in part by Grants EY03040 and EY01545 from theNational Institutes of Health.

Submitted for publication June 8, 1982.Reprint requests: Stephen J. Ryan, MD, Department of Oph-

thalmology, University of Southern California School of Medicine,Estelle Doheny Eye Foundation, 1355 San Pablo Street, Los An-geles, CA 90033.

cally with fluorescein angiography, and the lesionsdivided into active and regressive stages on the basisof permeability to fluorescein. Animals examined at5, 8, 9, 10,' 12, and 27 weeks after photocoagulationhad active leaking vessels seen on fluorescein angiog-raphy. One animal did not develop SRN, and a castpreparation was performed 11 weeks after photoco-agulation. Animals with previously active SRN, whichhad undergone involution, were studied by cast prep-aration 81, 82, and 158 weeks after laser photoco-agulation.

From 5 to 158 weeks after photocoagulation, theanimals were killed under sodium pentobarbital (30-70 mg) anesthesia. Both common carotid arterieswere cannulated with a 16-gauge plastic cannula. Thejugular veins were transected and perfusion was un-dertaken with 1500 cc physiologic saline and sodiumheparin (lOiu/cc) at a constant pressure of 140 mm/Hg. After the complete replacement of blood, a mix-ture of 60 cc methyl-methacrylate monomer (Mer-cox) and 1 % accelerator was perfused until the resinappeared at the jugular veins. The vessels in the neckwere then clamped until the resin hardened, whichusually took from 10 to 20 min. The eyes were enu-cleated and kept intact in 15% KOH at 37-40 C for3 weeks with daily changes of the KOH. After thistime, the cast was rinsed, air dried, and the entireposterior pole isolated under a dissecting microscope.The cast was mounted onto an SEM stub and coatedwith 30 nm of gold-palladium. The cast was thenexamined with a JEOL SEM 35 scanning electronmicroscope at an accelerating voltage of 10 or 15 KV(inclined angle 0°, 10°, or 15°, working distance 39mm). Following examination from the choroidal side,the cast was remounted, coated with an additional

0146-0404/8 3/0400/481/$ 1.30 © Association for Research in Vision and Ophthalmology

481

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933341/ on 05/11/2018

482 INVESTIGATIVE OPHTHALMOLOGY b VI5UAL SCIENCE / April 1983 Vol. 24

Fig. 1. Fluorescein angiogram PA weeks after onset of neovascularization (Monkey M5). Left, taken during early phase of angiography,site nos. 1, 2, 3, and 4 were studied with the scanning electron microscope. Right, late venous phase, leakage of dye is evident.

25 nm of gold-palladium, and examined from theretinal side.

Results

Active Leaky Stage

The eyes were examined 8 weeks after laser pho-tocoagulation (this was usually 4 to 5 weeks after thefluorescein angiogram suggested the presence ofSRN). Neovascular tufts were found at the site ofphotocoagulation that showed leakage on the fluo-rescein angiogram. Site nos. 1, 2, 3, and 4 (Fig. 1)were studied. The cast of the specimen obtained fromsite no. 1 showed a well-developed, branching neo-vascular network (Fig. 2A); these new vessels origi-nated from the choroidal vasculature and extendedinto the subretinal space. The afferent vessel (Fig. 2A,arrow A) was derived from the choroidal artery,which showed many imprints of endothelial nucleiand fine grooves. The efferent vessel (Fig. 2A, arrowE), collecting numerous venules, showed no obviousnuclear imprints on the surface of its replica; neitherwere nuclear imprints obvious at the tip of the SRNcapillary, probably due to the absence of smoothmuscle in the vessel walls. One plastic globule couldbe seen at the tip of the SRN frond (Fig. 2B, arrow),which may be due to the escape of resin.

Fluorescein angiogram of site no. 2 showed a smallsite of leakage at the edge of the photocoagulationsite (Fig. 1). The cast specimen of this lesion (Fig. 3)demonstrated a small neovascular network; the ves-sels feeding and draining this network could be vi-sualized in the area where fluorescein leaked andpooled. At the base of the neovascular network, oneof the two main vessels connected to the choriocap-

illaris (Fig. 3, arrow E). From the scleral side, anothermain vessel, presumably the afferent vessel, con-nected to the arterial branch of the choroidal artery(Fig. 4 A) and could be traced back to one of the shortposterior ciliary arteries (Fig. 4B). Thus, the patternof circulation at this site was from the posterior ciliaryartery to the frond, then draining into efferent venulesto the choriocapillaris.

Although no other sites showed marked leakage onfluorescein angiography, the cast preparation of siteno. 3 demonstrated a tiny, delicate neovascular tuftat the edge of the lasered site (Fig. 5). These newvessels were approximately equidistant from the ar-teriolar and choriocapillaris components. In addition,in site no. 4 some slender sprouts from the normalchoriocapillaris were seen at the edge of the burn site(Fig. 6). These sprouts are closely associated with thevenous side of the circulation.

Non-leaky Regressive Stage

Eighty-two weeks after photocoagulation, SRN(which had previously leaked fluorescein profusely)now failed to leak and pool fluorescein, suggestingthe lesion to be in regression (Fig. 7). The correspond-ing vascular cast showed a tuft with a glomerular ap-pearance, consisting of tightly packed vessels thatgrew over the SRN site and connected with anotherSRN lesion (Fig. 8A). Although its connection to thechoroidal vasculature was masked by the packed net-work of the capillaries in the right SRN site, the af-ferent (Fig. 8B, arrow A) and efferent (Fig. 8B, arrowE) vessels can be identified in the left SRN site. Itbecame obvious from the stereo pair pictures of theleft SRN site, disregarding the right SRN frond com-municating channel, that the circulation of this SRN


No. 4 SUDRETINAL NEOVASCULAR CASTS / Ohhumo and Ryan 483

.^ * >Jfcir

Fig. 2. A, Cast study performed 8 weeks after laser photocoagulation (Monkey M5). Retinal view of site no. I. The defect of methyl-methacrylate cast of the choriocapillaris (c) was the laser burn site, and SRN vessels were prominent around the defect. The afferent vessel(arrow A) obviously originated from the choroidal arterial branch, which showed numerous imprints on endothelial nuclei, and the efferentvessel (arrow E), which showed no obvious surface nuclear imprints, also could be observed (X120). B, Higher magnification of the inferiorSRN tips shown in Figure 2A. Note the plastic globule (arrow) at the tip of the SRN frond (X236).


484 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / April 1983 Vol. 24

Fig. 3. Retinal views of site no. 2 (Monkey M5).The afferent vessel (arrow A) terminated in theneovascular frond. The frond then drained via theefferent vessel (arrow E) as it coursed to (he cho-riocapillaris (X180).

Fig. 4. A, Scleral view of site no. 2 (Monkey M5). The afferentvessel (arrow A) connecting the branch of the choroidal artery(arrow Art) coursed to the rim of the burn site (XI46). B, Lowmagnification view of Figure 4A. The boxed area correspondsto site no. 2. Surrounding the optic nerve (N) note one of theshort posterior ciliary arteries (SPCA), after many branches,ultimately supplies the afferent vessel (arrows) to the neovas-cular frond (X20).


No. 4 SUDftETINAL NEOVA5CULAR CA5T5 / Ohhumo and Ryan 485

Fig. 5. Retinal view of siteno. 3 (Monkey M5). Thesimple SRN frond appearsin the lasered region. Ar-rows A and E show the af-ferent and efferent vesselsrespectively (X160).

was arterial through the frond and drained into ven-ules without going through the choriocapillaris.

In another example, fluorescein angiographyshowed that the SRN had undergone regression (Fig.9) 156 weeks after photocoagulation. Although thesevessels were not apparent on the fluorescein angio-gram, the vascular cast exhibited a persistent, al-though delicate, vascular loop. In the upper site, theafferent vessel (Fig. 10A, and arrow A) can be iden-tified, but the efferent vessel (Fig. 10A, arrow E) can-not be traced to its base. It may have broken duringthe cast processing or not injected well with the resin,therefore being invisible. In the bottom site, the af-ferent (Fig. 10B, arrow A) and efferent (Fig. 10B, ar-row E) vessels can be observed and traced at the baseof the vascular frond. It is probable that the smallluminal size resulted from decreased blood flowthrough the regressed vessels.

Discussion

The subretinal neovascularization and the deter-mination of the disciform response has been the sub-

ject of some clinical pathological correlation as notedby Gass,11 Sarks,12 Small, Green, Alpar, and Drewry,l3

and Green and Key.14 The material available forstudy is usually a result of a postmortem examination.It is recognized that subretinal neovascularizationand the disciform process may run its course overmany years as in senile macular degeneration withrepeated exudation and serous detachment, hemor-rhage, and cicatrization. Thus, subretinal neovascu-larization is a dynamic process with evolution andchange of appearance of the subretinal vessels. Theinitial leaking phase and the regressed or nonleakingphase may be the result of maturation or wound heal-ing response or may be related to the retinal com-partment and extracellular control of tight vesseljunctions. This is in contrast to the marked porosityof the choroidal compartment from which these ves-sels are initially derived. The current experimentalstudy was designed in part to study the evolution ofthis process over time and to determine the three-dimensional array and appearance of subretinal neo-vascularization.


486 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / April 1983 Vol. 24

Fig. 6. Retinal view of site no. 4 (Monkey M5). Small arrows indicate slender sprouts from the choriocapillaris. Note the close associationto the venule (large arrow) (X180).

Our morphologic studies have included ultrastruc- of the pathogenesis of this problem is the ignoranceture and horseradish peroxidase tracer analysis.2 One of the three-dimensional aspects of the extent andof the major problems related to the understanding nature of the subretinal neovascularization. The te-

Fig. 7. Left, fluorescein angiogram 19 weeks after onset of neovascularization. Right, same fundus 66 weeks later shows that this nolonger has active leakage and pooling of fluorescein.


No. 4 SUBRETINAL NEOVASCULAR CASTS / Ohkumo and Ryan 487

Fig. 8. A, Cast study performed 82 weeks after laser photocoagulation (Monkey M467). Cast preparation of SRN demonstrated in Figure7 (arrows). Now, 82 weeks after photocoagulation. Two glomerular configurations of new vessels overlying the choriocapillaris plane (c)(X45). B, Higher magnification of the boxed area in Figure 8A. Note the afferent (arrow A) vessels supplying the frond and the efferent(arrow E) vessels draining the frond and bypassing the choriocapillaris (X160).


468 INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE / April 1983 Vol. 24

Fig. 9. Left, fluorescein angiogram taken 2 weeks after the onset of neovascularization. Right, 94 weeks later of the same fundus withno leakage. Arrows indicate the sites observed under scanning electron microscope (see Figs. 10A,B).

dious techniques of serial reconstruction with ultra-structural correlation prevented a thorough study ofthe distribution and characteristics of these vesselsover time. The application of the perfusion cast tech-nique allows an excellent understanding of the char-acteristics and connections of these blood vessels.

As previously described,8 during the early stage ofthis study, we faced the problem of defective casting.However, we found that resin injections without per-fusion of a fixative provided an improved filling ofthe SRN vessels. After that, resin injection was per-formed on deeply anesthetized animals that had beenwell perfused with saline.

The plastic cast replicas help to bridge the gap be-tween histologic and fluorescein angiographic obser-vations of SRN. The results of these various types ofstudies are basically similar and show that the newlyformed vessels course from the lasered areas in theform of a neovascular frond. These plastic castsclearly demonstrate the afferent and efferent vessels.A number of possibilities of the origin of SRN canbe considered. In the earliest stage, a choriocapillaris-to-choriocapillaris shunt, an arteriole-to-choriocap-illaris shunt, or a choriocapillaris-to-venule shuntcould exist. A fourth possibility is that there could beSRN vessels directly linking an arteriole and venulewithout involvement of the choriocapillaris. Someearly specimens demonstrated slender blind-endedvessels at the edge of the lesion that originated fromthe choriocapillaris. Clark and Clark15 found neo-vascularization to occur at the capillary level, whileMichaelson,16'17 Wise,18 and Ashton'9 observed neo-vascularization to occur closer to the venule side. Weobserved the slender vessels from the venous side of

the choriocapillaris more frequently than from thearterial side.

We can speculate that the slender vessels representthe onset of neovascularization. We also suggest thatthe initiating event in the development of SRN maybe the occurrence of capillary buds around the laseredlesion on the venular side of the choriocapillaris; sub-sequently, arteriolar connections may form. As amatter of course, this process consists of reversal offlow through the parent choriocapillaris via the newarterial connection. Once the arteriole-choriocapil-laris-venule unit is established, the SRN frond maygradually increase in size and the arteriole-capillary-venule may gain access to the subretinal space as theSRN progresses to bypass the choriocapillaris system.If this hypothesis is correct, then pre-existing chorio-capillaris and its connection to an arterial branch ofthe choroidal artery appears to be necessary at theonset of subretinal neovascular proliferation. In apreliminary report,8 we speculated that a basic featureof SRN is its uniformity, ie, the afferent vessel is lo-cated near the center of the individual burn site, thencourses to the capillary net and the efferent vesselsfrom the net are connected to the choriocapillaris atthe periphery of the burn site. However, from thepresent long-term study, we discovered that as theSRN develops, a direct connection between the ef-ferent vessel and the choroidal venous system maybypass the choriocapillaris.

It is noteworthy that although few vessels are ap-parent by fluorescein angiography, cast preparationsbetter demonstrate such vessels. In fact in the latestages (Fig. 7, 8), extensive vessels, even in a glo-merulus formation, may be apparent on cast prepa-


No. 4 SUQRETINAL NEOVASCULAR CA5T5 / Ohkumo and Ryan 489

Fig. 10. A, Cast study performed 158 weeksafter laser photocoagulation (Monkey M248A).Retinal view of the upper site. The caliber of SRNvessels is less than that of the surrounding cho-riocapillaris. The afferent vessel (arrow A) arisesfrom the choroidal artery. However, the efferentvessel (arrow B) cannot be traced to its base(X100). B, In the bottom site, the afferent vessel(arrow A) can be traced to the choroidal arterialbranch (arrow Art). The efferent vessel (arrow E)can also be traced to the choroidal vein (XI00).

'&£$•ration despite the absence of their appearance on flu-orescein angiography. Even in the more late stage,the vessels also exist as thin, delicate, presumablyatrophic loops (Figs. 10A,B). As expected, the plasticcast technique is limited in that the injection pressureis not consistent throughout the smaller vessels. It ispossible that some distension of the vessels may leadto opening of endothelial junctions or breakdown ofendothelial walls as seen in Figure 2B. In spite of goodperfusion of the resin and thicker coating of gold-

palladium, we often encountered a problem of charg-ing during scanning electron microscopy. It is prob-ably due to the inherent poor connection of SRNfrond to choroidal circulation. Moreover, some ves-sels, particularly closed-end vascular sprouts, are notconsistently demonstrated by the injection of resin.Also, solid cords of SRN vessels and morphologicrelationships of SRN vessels to surrounding tissuescannot be examined. Accordingly, more data isneeded to explain the mechanism of the neovascu-


490 INVESTIGATIVE OPHTHALMOLOGY 6 VISUAL SCIENCE / April 1983 Vol. 24

larization. As already demonstrated by cast prepa-ration, the developing SRN presents a characteristicpattern. The pattern of the early immature capillarynetwork differs from that at a later stage; clearly thegrowing vessels undergo gross modification duringmaturation. The possible role of extracellular controlon the characteristics such as permeability of thesevessels derived from the choroidal circulation mustremain a subject for further study. We can speculatethat there may be a difference between the retinal andchoroidal compartments in terms of cellular or ex-tracellular control. We note also that absence of flu-orescein leakage in the regressive lesions results fromactual decrease of leakage rather than from completedisappearance of the abnormal vessels. These resultsmay help to explain some aspects of the dynamicprocess of subretinal neovascularization.

Key words: experimental subretinal neovascularization,rhesus monkey, methyl-methacrylate, vascular cast, scan-ning electron microscopy

References

1. Ryan SJ, Mittl RN, and Maumenee AE: The disciform re-sponse: An historical perspective. Albrecht von Graefes ArchKlin Exp Ophthalmol 215:1, 1980.

2. Ohkuma H and Ryan SJ: Permeability of experimental sub-retinal neovascularization in the monkey. Arch Ophthalmol,in press.

3. Matsuo N: Scanning electron microscopic studies on the cor-rosion casts of the choroidal blood vessels. In The Structureof the Eye. Ill, Yamada E and Mishima S, editors. Tokyo, JpnJ Ophthalmol, 1976, pp. 407-417.

4. Matsusaka T: Angio-architecture of the choroid. Jpn JOphthalmol 20:330, 1976.

5. Shimizu K and Ujiie K: Structure of Ocular Vessels. Tokyo,Igaku Shoin Ltd., 1978, pp. 54-55.

6. Uyama M, Itotagawa S, Fukumi K, and Doi H: Angioarchi-tecture of the choroid studied by plastic casts in the choroidalvasculature. Concilium Ophthalmologicum 23rd, 1978, Kyoto.2:1075, 1979.

7. Risco JM and Nopanitaya W: Ocular microcirculation; a scan-ning electron microscopy study. Invest Ophthalmol Vis Sci19:5, 1980.

8. Ohkuma H and Ryan SJ: Vascular casts of experimental sub-retinal neovascularization in monkeys: A preliminary report.Jpn J Ophthalmol 26:150, 1982.

9. Ryan SJ: The development of an experimental model of sub-retinal neovascularization in disciform macular degeneration.Trans Am Ophthalmol Soc 77:707, 1979.

10. Ryan SJ: Subretinal neovascularization after argon laser pho-tocoagulation. Albrecht von Graefes Arch Klin Exp Ophthal-mol 215:29, 1980.

11. Gass JDM: Pathogenesis of disciform detachment of the neu-roepithelium. Am J Ophthalmol 63:573, 1967.

12. Sarks SH: New vessel formation beneath the retinal pigmentepithelium in senile eyes. Br J Ophthalmol 57:951, 1973.

13. Small ML, Green WR, Alpar JJ, and Drewry RE Jr: Senilemacular degeneration; a clinicopathologic correlation of twocases with neovascularization beneath the retinal pigment ep-ithelium. Arch Ophthalmol 94:601, 1976.

14. Green WR and Key SN III: Senile macular degeneration: Ahistopathologic study. Trans Am Ophthalmol Soc 75:180,1977.

15. Clark ER and Clark EL: Microscopic observations on thegrowth of blood capillaries in the living mammal. Am J Anat64:251, 1939.

16. Michaelson IC: The mode of development of the vascular sys-tem of the retina, with some observations on its significancefor certain retinal diseases. Trans Ophthalmol Soc UK 68:137,1948.

17. Michaelson IC: Vascular morphogenesis in the retina of thecat. J Anat 82:167, 1948.

18. Wise GN: Factors influencing retinal new vessel formation.Am J Ophthalmol 52:637, 1961.

19. Ashton N: Oxygen and the growth and development of retinalvessels; In vivo and in vitro studies. Am J Ophthalmol 62:412,1966.


vascular casts of experimental subretinal neovascularization in...

Documents