Spontaneous Corneal Neovascularization in Nude MiceLocal Imbalance Between Angiogenic andAnti-angiogenic Factors

Grazyna M. Kaminska and Jerry Y. Niederkorn

Purpose. The present study considered the hypothesis that spontaneous neovascularization ofthe cornea of nude mice results from a local imbalance between angiogenic and anti-angio-genic factors.

Methods. The presence of angiogenic and anti-angiogenic factors was revealed by exchangingorthotopic corneal allografts between nude BALB/c mice and normal (hirsute) euthymicBALB/c mice and observing the presence, intensity, and degree of corneal neovascularizationbefore and after grafting.

Results. Avascular corneal grafts from normal BALB/c donors resisted neovascularizationafter grafting to spontaneously vascularized graft beds in nude mice. In contrast, spontane-ously vascularized corneal grafts from nude mice remained vascularized over 2 mo after graft-ing to similar nude recipients. Although corneal grafts from nude donors stimulated neovascu-larization in normal BALB/c recipients, most of the vessels regressed by the 6th week posttransplantation.

Conclusions. The results confirm that there is an imbalance between angiogenic and anti-angio-genic factors in the nude mouse cornea. The cornea of the nude mouse displays more angio-genic activity and less anti-angiogenic activity than that of the normal mouse. Most angiogenicactivity of the nude mouse cornea appears to reside in the epithelium. Invest Ophthalmol VisSci. 1993;34:222-230.

1 he normal cornea is avascular but may be invadedby blood vessels in a variety of pathologic conditions.1

Avascularity of a cornea was suggested to depend onthe presence of anti-angiogenic factors and on the

From the Department of Ophthalmology, The University of TexasSouthwestern Medical Center at Dallas, Dallas, Texas.Supported in part by National institutes of Health Grant EY07641, andby an unrestricted grant from Research to Prevent Blindness, Inc., NewYork, New York. Dr. Niederkorn is a Research to Prevent Blindness, Inc.,Senior Scientific investigator.Submitted for publication: February 26, 1992; accepted June 26, 1992.Proprietary interest category: N.Reprint requests: Jeny Y. Niederkorn, Department of Ophthalmology,University of Texas Southwestern Medical Center, 5323 Hairy HinesBlvd., Dallas, TX 75235-9057.

properties of a corneal stroma that form a barrieragainst penetrating blood vessels.23 Thus, corneal neo-vascularization would be caused by a decreased con-centration of anti-angiogenic factors, or by a loosen-ing of the stroma during inflammatory processes. An-other important cause of the in-growth of bloodvessels into cornea may be a local increase in angio-genic factors that would attract and stimulate prolifer-ation of blood vessels. Such factors may originate frominflammatory cells that invade the cornea4"6 or fromthe cornea itself,7 and their over production mightcontribute to corneal neovascularization.

In our laboratory, a new model of spontaneouscorneal neovascularization has been described re-

222Investigative Ophthalmology & Visual Science, January 1993, Vol. 34, No. 1Copyright © Association for Research in Vision and Ophthalmology

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

Spontaneous Corneal Neovascularization in Nude Mice 223

cently in nude and hairless mice.8 We noted that thesemice show corneal neovascularization that starts in theperinatal period and progresses without a consistentinflammatory infiltrate or other apparent stimuli.These vessels persist without any sign of regression.

The aim of the present study was to test the hy-pothesis that this spontaneous corneal neovasculariza-tion results from a local imbalance between angio-genic and anti-angiogenic factors.

To reveal the presence of such factors and localizethem in different layers of the cornea, we orthotopi-cally transplanted full-thickness corneal buttons, aswell as fragments of corneal epithelium and cornealstroma, and assessed the vascularity of the grafts andthe host cornea using spontaneously vascularizingBALB/c nude and normal BALB/c mice in variousdonor-recipient combinations.

We found that corneas of the nude mice had muchmore angiogenic activity than those of normal miceand that this activity was localized in the epithelium.We also found that corneal stroma in nude mice, incontrast to normal mice, lacked anti-angiogenicactivity.

MATERIALS AND METHODS

Mice

Female BALB/c and BALB/c nude mice (nu/nu) 8-10wk old were obtained from Simonsen Laboratories(Gilroy, CA) and kept at UTSMC in Dallas in a laminarair flow room with filtered air and sterile equipment,food and water. The animals were used for experi-ments within 5-7 days after arrival. All animals werehandled in accordance with the National Institutes ofHealth "Guide for the Care and use of LaboratoryAnimals" and the ARVO Resolution on the Use ofAnimals in Research. All studies adhered to the tenetsof the Declaration of Helsinki.

Corneal Grafting

Anesthesia. General anesthesia was achieved withan intraperitoneal injection of sodium pentobarbital(1-2 mg/mouse; Abbott Laboratories, North Chicago,IL). For topical anesthesia, proparacaine hydrochlo-ride (Alcon Laboratories Inc., Puerto Rico) was ad-ministered, and to dilate pupils, tropicamide 1% (Al-con) was used.

Full-Thickness Orthotopic Corneal Grafts. The pro-cedure was modified from that described previously.9

Briefly, a full-thickness (1.5 mm diameter) penetratingcorneal button was obtained from the donor using a1.5 mm trephine and vannas scissors. The recipientcornea was similarly scored with a 1.5 mm trephine,

and the central 1.5 mm button was removed. The do-nor graft was sewn in place using seven interruptedsutures with 11-0 nylon (Alcon Laboratories Inc., Ft.Worth, TX). The anterior chamber was restored byinjecting 5-10 /i\ sodium hyaluronate (Healon; AlconLaboratories Inc.). Sutures were removed 10 dayslater. Topical Polysporin ophthalmic antibiotic oint-ment (Burroughs Wellcome Co, Research TrianglePark, NC) was applied after surgery, and every 2 daysfor 10 days. The eyelids were closed for 24 hr with asingle suture 7-0 silk (Ethicon Inc., Johnson & John-son Co., Somerville, NJ). Eyelids were re-opened 24 hrlater by removing the sutures.

Epithelium-Depleted Orthotopic Corneal Grafts. Do-nor cornea was depleted of epithelium by scrapingwith a cotton swab under a slit-lamp microscope (theefficacy of this procedure was confirmed histologicallyin pilot experiments). A corneal button (containingstroma and endothelium) was procured and trans-planted orthotopically as described above.

Corneal Pocket Assay. The intracorneal pocket as-say was modified after that described by Muthukka-ruppan and Auerbach.10 A central cornea was incisedwith a Seamless Steri-Sharps No. 15 surgical blade(Seamless Co., Wallinford, CT) to obtain a vertical in-cision 0.5 mm long that penetrated approximatelyone-half through the cornea. A corneal pocket thenwas made by inserting the surgical blade into the inci-sion and carefully extending it up to 1 mm from thelimbus. Full-thickness or epithelium-depleted cornealtissue fragments (approximately 0.5 mm in diameter)or scraped corneal epithelium from a single donorthen were inserted into the pocket. The pocket fas-tened up spontaneously and healed in 24 hr. TopicalPolysporin antibiotic ointment was applied after theprocedure. The surgical procedure to form a pocketonly (sham operation) did not induce vascularization.

All surgical procedures were done under sterileconditions, using a Zeiss (Thornwood, NY) slit-lampmicroscope.

Experimental Design. Cross transplantation of cor-neal tissue between normal BALB/c and BALB/c (nu/nu) mice that had spontaneously vascularized corneaswas done in all four possible donor-recipient pair com-binations.

Clinical ObservationGrafted mice were observed with a slit-lamp micro-scope twice a week throughout the 2 mo study period.Graft opacity and edema were scored as described pre-viously.11 The progress of vascularization of grafts wasdocumented by counting the vessels growing towardthe corneal graft and by examining vasodilation in thelimbal region. The number, tortuosity, directionality,length, and caliber of blood vessels were noted.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

224

Statistical Evaluation

A statistical evaluation of the results was done at Aca-demic Computing Services at UTSMC with the help ofDr. D. D. Mclntire. A three-factor full factorial designwas used to compare the number of vessels per graft asa response to orthotopic corneal grafts. The factors(donor, recipient, and type of the graft) each were pre-sented at two levels: (1) BALB/c and nude for eachdonor and recipient; and (2) full thickness and epithe-lium-depleted grafts as the two levels of graft type. Tenmice were randomized to each of the eight possibletreatment combinations (2 donor X 2 recipient X 2graft).

In studying the response to cornea fragments, a 2donor by 2 recipient by 3 graft full factorial design wasthe selected approach. Each donor and recipient werepresented at two levels—BALB/c and nude. The threelevels for graft were full-thickness, without epithelium,and epithelium only. Statistical comparisons amongthe means were accomplished using analysis of vari-ance of the full factorial design with two-factor inter-actions. Effects were considered statistically signifi-cant when the A* value was less than 0.05.

Histologic Examination

After the animals were killed with methoxyHurane(Pitman-Moore Co., Washington Crossing, NJ), mouseeyes were collected and fixed in 10% phosphate-buffered formalin. The specimens were embedded inparaffin, sectioned, and stained with hematoxylin-eosin.

Investigative Ophthalmology & Visual Science, January 1993, Vol. 34, No. 1

Fragments of cartilage were processed similarly toexamine whether spontaneous vascularization in nudemice affects other avascular tissues.

RESULTS

Morphology of Nude Mouse Cornea

Corneas of BALB/c (nu/nu) mice showed extensivevascularization by blood vessels that penetrated fromthe limbus into the stroma. These vessels extendedclosely to the center of the cornea, leaving only a tinycentral part nonvascularized. Some of them formedloops and returned to the limbus. A majority of thevessels had a rather straight course and showed few, ifany bifurcations. On hislologic preparations, bloodvessels were located in the superficial stroma in thevicinity of the epithelial layer (Fig. 1).

Orthotopic Full-Thickness and Epithelium-Depleted Corneal Grafts

Normal BALB/c to BALB/c Donor-Recipient Combi-nation. The neovascularization elicited by syngeneicfull-thickness grafts was not noted until day 5 (Fig. 2)and persisted up to 10 days, reflecting a nonspecificresponse to corneal injury. The blood vessels origi-nated from the limbus and reached the graft, but didnot penetrate it. These vessels were of medium caliberand ran a straight course without loop formation.They started to regress on the 10th day after trans-plantation, when sutures were removed. Most of theblood vessels around the suture placement sites re-gressed within 4 days after suture removal, but occa-

«*Mf****\&x

FIGURE 1. Constitutive vascu-larization of normal nudemouse cornea. Large bloodvessels (arrows) in superficialstroma of the central corneaappear to be capillaries andpost-capillary venules. (Hema-toxylin-eosin, original magnifi-cation X500.)

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

Spontaneous Corneal Neovascularization in Nude Mice 225

FIGURE 2. BALB/c orthotopic corneal graft on a syngeneic BALB/c recipient. Photographytaken 3 days post-transplantation. Note avascularity of the graft.

sionally some of the small vessels persisted throughoutthe study period. Generally, in the control syngeneicgroup, there were no blood vessels in the graft ob-served by the 14th day post-transplantation, exceptthose few related to suture placement.

Initial, nonspecific vascular response to epithe-lium-depleted orthotopic corneal grafts was less in-tense than the response to full-thickness grafts. Novessels that penetrated the graft were seen after 14days.

Nude BALB/c to Nude BALB/c Donor-RecipientCombination. Vascularized full-thickness grafts notonly retained their vascularity, but gained new bloodvessels from the host limbus. These new vessels were ofmedium caliber and were straight and without bifurca-tions. Generally, nude to nude corneal grafts stayedvery well vascularized over the entire 2 mo observationperiod. The vascular response to epithelium-depletedgrafts was lower than in full-thickness grafts, but morepronounced than in normal syngeneic BALB/c recipi-ents.

Nude BALB/c to Normal BALB/c Donor-RecipientCombination. Vascularized full-thickness corneal graftsstimulated the growth of host vessels from the limbusthat resulted in additional graft vascularization (Fig.3). However, most of the donor and recipient vesselsstarted to regress around the 6th wk after transplanta-tion; by 2 mo, no blood vessels could be seen in thegrafts, and only "ghost" vessels were noted on histo-logic preparations.

Epithelium-depleted grafts induced less vascularresponse than full-thickness grafts, with similar re-gression of blood vessels occurring by 6 wk.

Normal BALB/c to Nude BALB/c Donor-RecipientCombination. Corneal full-thickness grafts trans-planted onto vascularized recipient cornea remainedavascular over the entire observation period. Initialhost vascular response to such grafts was similar tothat observed in syngeneic normal controls, with sev-eral blood vessels running toward the graft. However,a striking feature was that these vessels never pene-trated the graft. They usually were able to reach thegraft border, but then changed their direction and en-circled the graft. Epithelium-depleted grafts did notinduce significant vascular response.

The results of the full-thickness corneal graft ex-periments are summarized in Table 1.

Corneal Pocket Grafts

Normal BALB/c to BALB/c. Full-thickness and epi-thelium-depleted corneal fragments (0.5 mm diame-ter) and isolated corneal epithelium failed to induceany vascular response in the recipient corneas.

Nude BALB/c to Nude BALB/c. All types of graftsinduced a strong vascular response in the already va-scularized recipient corneas, but the intensity of thisresponse was highest for full-thickness fragmentgrafts. The number of blood vessels always increased,but no changes in the vascularization pattern were ob-served.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

Investigative Ophthalmology & Visual Science, January 1993, Vol. 34, No, 1

FIGURE 3. Corneal graft from nude BALB/c donor grafted to a normal BALB/c recipient.Photograph taken 27 days post-transplantation. Note intense neovascularization of entirecorneal graft.

Nude BALB/c to Normal BALB/c. Only full-thick-ness and isolated corneal epithelium grafts elicited thevascular response in the recipient corneas. This re-sponse was weak, and blood vessels barely reached thegrafts.

Normal BALB/c to Nude BALB/c. Transplanted tis-sues were not vascuiarized by host vessels. The vesselsthat normally are present in the nude mouse corneaseemed to be repelled from the grafts of normalBALB/c corneal fragments (Fig. 4) but not of corneal

TABLE l. Vascular Response to OrthotopicCorneal Grafts, Noted 2 wk AfterTransplantation

Donor

BALB/cNudeNudeBALB/c

Recipient

BALB/cNudeBALB/cNude

Number of New

Full-ThicknessGraft

5 (3-6)23 (20-33)18 (14-28)6 (3-8)

Vessels Per Graft*

Epithelium-DepletedGraft

4 (2-5)17(11-19)11 (8-12)6 (4-7)

* Results are means of newly formed blood vessels per graft withranges in parentheses. Ten mice per group.The statistical evaluation showed the two significant interactions:donor X recipient (P = 0.0037) and donor X graft (P = 0.0001). Inboth cases, the nude donors provided a greater number of vesselsper graft. The full-thickness nude cornea grafts exhibited a greatermean of vessels than the epithelium-depleted grafts.

epithelium alone. The results of the pocket graft ex-periments are summarized in Table 2.

Histology of Nude Hyaline Cartilage. Cartilage innude mice was avascular, as was cartilage in normalBALB/c mice (Fig 5). A typical hyaline cartilage matrixwith embedded chondrocytes lying singly or in a fewisogeneic groups was noted. The matrix was free ofany blood vessels.

DISCUSSION

The results indicate there is an imbalance between an-giogenic and anti-angiogenic factors in the spontane-ously vascuiarized corneas of the nude mice. The cor-nea of the nude mouse displays more angiogenic activ-ity than that of a normal mouse. Full-thickness cornealgrafts from nude donors elicited a massive, persistentin-growth of blood vessels from the limbus in normalrecipients. In contrast, similar grafts from normalmice failed to induce any significant, persistent angio-genic responses. The angiogenic activity was localizedin the epithelium, because isolated nude corneal epi-thelium of a total volume not exceeding 0.2 mm3 in-duced a marked angiogenic response after implanta-tion into a corneal pocket in normal mice. This re-sponse was, however, not as massive as a full-thicknessgraft. This difference might be explained by the muchsmaller volume of the epithelial implant, compared tothe full-sized corneal grafts. We found that even epi-

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

Spontaneous Corneal Neovascularization in Nude Mice 227

FIGURE 4. Normal BALB/c full-thickness corneal fragment 14 days after transplantation intoa corneal pocket in BALB/c nude mouse. Note the relative avascularity of the graft (arrow-head). Blood vessels in the superficial siroma of the nude mouse cornea are faintly visible.There is a single large caliber blood vessel (arrow) near the graft that seems to have beenrepelled from the graft.

thelium-depleted nude corneal grafts were weakly an-giogenic. We speculate that this was due to an accumu-lation of a putative angiogenic factor that diffused intothe stroma from the corneal epithelium. The concen-tration of such a factor probably was low and was notsustained because of the absence of corneal epithe-lium. Further support for the localization of putativeangiogenic factors in the epithelium is suggested bythe histologic observation that most of the blood ves-sels in the nude corneas were near epithelium.

In contrast to nude corneal grafts, isolated epithe-lial fragments from normal mice elicited a weak and

transient vascular response in the normal host, as didnormal full-thickness corneal grafts. Epithelium-de-pleted grafts failed to induce any neovascularization.

We also found that the corneal stroma in nudemice, in contrast to normal mice, showed diminishedanti-angiogenic activity.

Epithelium-depleted corneal grafts from normaldonors seemed resistant to vascular invasion whentransplanted into nude mice. Although some hostblood vessels initially were directed toward suchgrafts, they never penetrated them, and instead by-passed the grafts. This encircling movement seemed to

TABLE 2. Vascular Response to Corneal Fragments (0.5 mmdiameter) Implanted Into Corneal Pocket

Donor

BALB/cNudeNudeBALB/c

Recipient

BALB/cNudeBALB/cNude

Number of New Vessels Per

Full-Thickness

09 (5-10)4 (2-5)0

w/o Epithelium

05 (4-6)2 (1-4)0

Graft*

Epithelium Only

03 (1-4)2 (1-4)0

* Results are means of newly formed blood vessels per graft (ranges in parentheses) noted 2 wk aftertransplantation. Five mice per group.The statistical evaluation showed a significant interaction of recipient X graft (P = 0.0103). Vascularityof grafts from nude donors (full-thickness and \v/o epithelium) increased significantly in nude recipientscompared to BALB/c recipients.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

228 Investigative Ophthalmology 8c Visual Science, January 1993, Vol. 34, No. 1

x.

* • - • * .

be caused by the presence of a repelling substance thatprevented vessel penetration. Similar morphologicevents occur when blood vessels induced by a tumor orlymphocytes interact with anti-angiogenic substancesfrom cartilage.12-13

When nude epithelium-depleted corneal graftscontaining pre-existing blood vessels were trans-planted to nude mice, they were penetrated by addi-tional new blood vessels and remained vascularizedover a 2 mo observation period. However, when theywere transplanted into normal corneas, their stromalvessels regressed within 2 mo, probably because of aninsufficient angiogenic stimulus, which would beneeded to sustain the presence of vessels, or becauseof the presence of anti-angiogenic factors in the nor-mal host cornea.

Taken together, our results seem to suggest that inthe nude mouse cornea there is more of an imbalancein angiogenic factors (most probably epithelium de-rived) than in corneas of normal mice. At the sametime, the corneal stroma in the nude mice seems tolack anti-angiogenic factors that are present in thecorneas of normal mice.

It recently was proposed that spontaneous vascu-larization of the corneas in the nude and hairless mu-tant mouse strains results from chronic inflammationin response to persistent bacterial infections of theconjunctival cul-de-sac.14 Although we often have ob-served bedding and other foreign debris in the con-junctival cul-de-sac of nude and hairless mice, we havenot found convincing evidence to support the afore-mentioned hypothesis. Histologic examination (lightmicroscopy) of corneas collected from nude and hair-

FIGURE 5. Cartilage from thexiphoid process of a BALB/cnude mouse. Note avascularityof cartilage. Specimens col-lected from hairless (SKH1:hr/hr) mice also were avas-cular. Arrows indicate bloodvessels outside of cartilage ma-trix. (Hematoxylin-eosin, origi-nal magnification X500.)

less mice on days 4, 6, 8,13, 14, 18, 21, 23, 24, 28, 30,44, 49, 52, and 56 post-partum failed to establish atemporal relationship between the presence of poly-morphonuclear neutrophils (PMNs) and cornealblood vessels (unpublished findings). Some corneascontained numerous PMNs, yet were free of cornealvessels, whereas other corneas (especially those ofadult mice) contained vessels throughout the entirecornea, yet were free of PMNs. Although a significantnumber of corneas did contain extensive PMN inflam-mation and blood vessels, we are reluctant to concludethat the presence of corneal vascularization in nudeand hairless mice is a direct result of chronic inflamma-tion, because the association between vessels andPMNs is very inconsistent. Moreover, the inherent an-giogenic activity of nude corneal grafts transplanted tohirsute BALB/c recipients and the anti-angiogenic ac-tivity of BALB/c corneas grafted to nude recipientsfurther argues against the proposition that corneal vas-cularization in nude and hairless mutant mouse strainssimply is the product of chronic inflammation. None-theless, this issue warrants further scrutiny and couldbe resolved by a carefully controlled prospectivestudy.

The vascularization in nude mice seemed to affectonly corneas, because histologic observations of otheravascular tissues, such as cartilage, revealed normalstructure with no blood vessels. Another possibility forspontaneous neovascularization could be an excessivesusceptibility of nude mouse vascular endothelium toangiogenic stimuli. The in vivo reactivity of nude vascu-lar endothelium to angiogenic stimuli from tumor andlymphoid cells was tested in vivo using a modified lym-

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

Spontaneous Corneal Neovascularization in Nude Mice 229

phocyte- or tumor cell-induced angiogenesis assay15 inwhich intradermal injection of appropriate angiogeniceffector cells induces a cell-dose dependent angio-genic reaction. Using allogenic spleen and lymph nodecells, as well as UV5C25 tumor cells, we observed nosignificant differences in the angiogenic response be-tween BALB/c (nu/nu) and normal BALB/c mice (un-published findings).

It is interesting that spontaneous corneal neovas-cularization also has been observed in SKHl:hr/hrhairless mice.8 Because hairless mice are competentimmunologically, the defect of the immune systemdoes not seem to be related to the observed cornealanomaly.

All of this leads us to speculate that the majorcause of the observed spontaneous corneal neovascu-larization in the nude and hairless mutant mousestrains is a local imbalance between angiogenic andanti-angiogenic factors.

We do not know the nature of the putative angio-genic factors in the nude mouse cornea. Such factorsmay be produced by the corneal epithelium7 or byother tissues and stored in the cornea. Some of theangiogenic factors that bind to heparin—ie, fibroblastgrowth factor16—were suggested to be sequestered inDescemet's membrane and produce corneal vascular-ization upon their release.1718 Glycosaminoglycans inthe corneal stroma, such as heparin and hyaluronatefragments, also may be angiogenic.19 Thus, increasedangiogenic activity in nude corneas may be the resultof an increased production, storage, or release of or adecreased catabolism of angiogenic factors. It is possi-ble that increased bacterial infection in the eyes ofnude mice directly or indirectly modify some of theseprocesses.14

In addition to an increased angiogenic activity innude cornea, there is a concomitant decrease in anti-angiogenic activity. It may result from the decreasedproduction or storage of or the excessive destructionof anti-angiogenic substances. The presence of anti-angiogenic factors in the corneal stroma has been sug-gested,20 but nothing is known about their nature andrelation to other anti-angiogenic factors from otheravascular tissues such as vitreous21 or cartilage.I2>13-22

That anti-angiogenic factors work in concert with an-giogenic factors to control the growth, persistence, orregression of blood vessels has been shown in the car-tilage, in which a balance between angiogenic andanti-angiogenic factors1213 is responsible for the avas-cularity.

In conclusion, the present results suggest thatspontaneous corneal neovascularization in nude miceprobably is related to an excess of angiogenic factorsand a deficiency of anti-angiogenic factors. The bal-ance between angiogenic factors (mainly from corneal

epithelium7) and anti-angiogenic factors from cornealstroma20 thus may be crucial for maintaining the avas-cular state of a normal cornea.

This model of spontaneous corneal neovascular-ization in nude mice may help gain a better under-standing of the mechanisms that control the in-growth, persistence, and regression of blood vesselsinto the cornea.

Key Words

angiogenic factors, anti-angiogenic factors, cornea, neovas-cularization, nude mice.

A cknowledgments

The authors thank Dr. YuGuang He for instruction in thecorneal grafting technique; Ms. Marsha Pidherney and Ms.Jessamee Mellon for technical assistance; Mr. John Hornafor photographic assistance; and Ms. Pat Clarke for prepar-ing the manuscript.

References

1. KJintworth GK. Corneal Angiogenesis. A ComprehensiveCritical Review. New York: Springer-Verlag; 1991:1-135.

2. Kuettner KE, Croxen RL, Eisenstein R, Sorgente N.Proteinase inhibitor activity in connective tissue. Ex-perientia. 1974; 30:595-597.

3. Mann I, Pirie A, Pullinger BD. An experimental andclinical study of the reaction of the anterior segmentof the eye to chemical injury with special reference tochemical warfare agents. BrJ Ophthalmol. 1948; 13:5-171.

4. Fromer CH, KJintworth GK. An evaluation of the roleof leukocytes in pathogenesis of experimentally in-duced corneal vascularization. I. Comparison of ex-perimental model of corneal vascularization. Am JPathol. 1975; 79:537-554.

5. Polverini PJ, Cotran RS, Gimbrone MA, Unanue ER.Activated macrophages induce vascular proliferation.Nature. 1977; 269:804-806.

6. Epstein RJ, Stulting RD. Corneal neovascularizationinduced by stimulated lymphocytes in inbred mice. In-vest Ophthalmol Vis Sci. 1987; 28:1505-1513.

7. EliasonJA, Elliott JP. Proliferation of vascular endo-thelial cells stimulated in vitro by corneal epithelium.Invest Ophthalmol Vis Sci. 1987; 28:1963-1969.

8. Niederkorn JY, UbelakerJE, Martin JM. Vasculariza-tion of corneas of hairless mutant mice. Invest Ophthal-mol Vis Sci. 1990;31:948-953.

9. He YG, RossJ, Niederkorn JY. Promotion of murineorthotopic corneal allograft survival by systemic ad-ministration of anti-CD4 monoclonal antibody. InvestOphthalmol Vis Sci. 1991;32:2723-2728.

10. Muthukkaruppan VR, Auerbach R. Angiogenesis inthe mouse cornea. Science. 1979;205:1416—1418.

11. Callanan DG, Peeler JS, Niederkorn JY. Characteris-tics of rejection of orthotopic corneal allografts inrats. Transplantation. 1988;45:437-443.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018

230 Investigative Ophthalmology & Visual Science, January 1993, Vol. 34, No. 1

12. Brem H, Folkman J. Inhibition of tumor angiogenesismediated by cartilage. JExpMed. 1975; 141:427-439.

13. Kaminski M, Kaminska G, Jakobisiak M, Brzezinski W.Inhibition of lymphocyte-induced angiogenesis by iso-lated chondrocytes. Nature. 1977; 268:238-240.

14. Lee HWH, Klintworth GK. An evaluation of spontane-ously developing corneal angiogenesis in nude (nu/nu) and hairless (hr/hr) mice (abstract). Invest Ophthal-mol VisSci. 1992;33(suppl):777.

15. Sidky YA, Auerbach R. Lymphocyte-induced angio-genesis: A quantitative and sensitive assay of the graft -versus-host reaction. J Exp Med. 1975; 141:1084-1100.

16. Gospodarowicz D, Neufeld G, Schweigerer L. Molecu-lar and biological characterization of fibroblastgrowth factor, an angiogenic factor which also con-trols the proliferation and differentiation of meso-derm and neuroectoderm derived cells. Cell Differen-tiation 1986; 19:1-17.

17. Folkman J, Klagsbrun M, Sasse J, Wadzinski M,Ingber D, Vlodavsky I. Storage of heparin-binding an-giogenic factor in the cornea: A new mechanism for

corneal neovascularization (abstract). Invest Ophthal-mol VisSci. 1987;28(suppl):230.

18. Soubrane G, Jerdan J, Karpouzas I, et al. Binding ofbasic fibroblast growth factor to normal and neova-scularized rabbit cornea. Invest Ophthalmol Vis Sci.1990;31:323-333.

19. McAuslan BR, Reilly WG, Hannan GN, Gole GA. An-giogenic factors and their assays: Activity of formylmethionyl leucyl phenylalanine, adenosine diphos-phate, heparin, copper, and bovine endothelium stim-ulating factor. Microvasc Res. 1983;26:323-338.

20. Kaminski M, Kaminska G. Inhibition of lymphocyte-induced angiogenesis by enzymatically isolated rabbitcornea cells. Arch Immunol Ther Exp. 1978; 26:1075-1078.

21. Preis I, Langer R, Brem H, Folkman J, Patz A. Inhibi-tion of neovascularization by an extract derived fromvitreous. Am] Ophthalmol. 1977:84:323-328.

22. Moses MA, Sudhalter J, Langer R. Identification of aninhibitor of neovascularization from cartilage. Science.1990;248:1408-1410.

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933395/ on 06/24/2018