Inhibition of Rat Corneal Angiogenesis by 16-kDa Prolactin and by Endogenous Prolactin-like Molecules

Abstract

purpose. The cornea is an avascular organ, where induction of new blood vessels involves the turn-on of proangiogenic factors and/or the turn-off of antiangiogenic regulators. Prolactin (PRL) fragments of 14 kDa and 16 kDa bind to endothelial cell receptors and inhibit angiogenesis. This study was designed to determine whether antiangiogenic PRL-like molecules are involved in cornea avascularity.

methods. Sixteen-kDa PRL and basic fibroblast growth factor (bFGF) or anti-PRL antibodies were placed into rat cornea micropockets and neovascularization evaluated by the optical density associated with capillaries stained by the peroxidase reaction and by the number of vessels growing into the implants. Prolactin receptors in corneal epithelium were investigated by immunocytochemistry.

results. bFGF induced a dose-dependent stimulation of corneal neovascularization. This effect was inhibited by coadministration of 16-kDa PRL, as indicated by a 65% reduction in vessel density and a 50% decrement in the incidence of angiogenic responses. Corneal angiogenic reactions of different intensities were induced by implantation of polyclonal and monoclonal anti-PRL antibodies. Corneal epithelial cells were labeled by several anti-PRL receptor monoclonal antibodies.

conclusions. These findings show that exogenous 16-kDa PRL inhibits bFGF-induced corneal neovascularization and suggest that PRL-like molecules with antiangiogenic actions function in the cornea. PRL receptors in the corneal epithelium may imply that PRL in the cornea derives from lacrimal PRL internalized through an intracellular pathway. These observations are consistent with the notion that members of the PRL family are potential regulators of corneal angiogenesis.

Angiogenesis, the formation of new blood vessels is a key element for growth and development, reproduction, wound healing, and tissue repair. Angiogenesis is also a primary event in several pathologic processes, including tumor development, chronic inflammatory diseases, and various retinopathies. ^{1 2}  

Several stimulators and inhibitors of angiogenesis have been identified that modulate the ability of endothelial cells to digest the basement membrane and proliferate, migrate, and/or associate into a new capillary network. ^{3 4} The balance between naturally occurring inducers and inhibitors of angiogenesis determines the active neovascularization observed during embryogenesis as well as the vascular quiescence maintained by most tissues in adult life. ^{4 5} The cornea and the cartilage are normally avascular. Their resistance to vascular invasion appears to involve a predominance of inhibitors over stimulators of angiogenesis. ^{4 5 6}  

Among a long list of active peptides, members of the prolactin (PRL) hormonal family have been proposed as potential regulators of angiogenesis. ⁷ Fragments of PRL of 14 kDa and 16 kDa bind to endothelial cell receptors and inhibit endothelial cell proliferation, migration, and tube formation. ^{7 8 9 10 11 12} Similarly, the 16-kDa amino terminal fragment of PRL inhibits the in vivo development of the microvasculature of the chick chorioallantoic membrane (CAM). ¹⁰ In addition proliferin and proliferin-related protein (PRP), considered members of the PRL family on the basis of primary sequence homology, compete with 16-kDa PRL for endothelial cell binding ⁹ and have opposite effects on angiogenesis. ¹³ Proliferin stimulates and PRP inhibits endothelial cell migration and corneal neovascularization, respectively. ¹³  

The cornea is a powerful model for investigating the in vivo regulation of blood vessel growth. ¹⁴ In the present study, we used rat corneas to analyze the potential contribution of PRLs to the regulation of angiogenesis. Our findings extend those previously reported of in vivo antiangiogenic action of 16-kDa PRL ¹⁰ and provide evidence to suggest that PRL-like molecules with inhibitory actions on blood vessel growth, may be involved in the maintenance of corneal avascularity. A preliminary report of these findings has been presented. ¹⁵ In all facets of the study, animals were managed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Materials and Methods

Reagents

Rat 23-kDa PRL (BIO grade) was donated by the National Hormone and Pituitary Program (NHPP, Torrance, CA). Sixteen-kDa PRL was generated after the enzymatic proteolysis of rat 23-kDa PRL with a particulate fraction from rat mammary gland extracts, gel filtration, and carbamidomethylation, as reported. ¹⁶ Human basic fibroblast growth factor (bFGF) was a kind gift from Judith A. Abraham (Scios, Mountain View, CA). Antisera were raised in rabbits by immunization with NHPP rat 23-kDa PRL or with rat 16-kDa PRL standards. Rat PRL monoclonal antibody (INN-rPRL-1) in raw ascites fluid, generated and characterized as described, was used. ¹⁷ Four monoclonal antibodies (U5, U6, T1, and T6) directed against the extracellular domain of the PRL receptor were kindly provided by Paul A. Kelly (Institut National de la Santé et de la Recherche Médicale U-344, Paris, France). Two of these monoclonal antibodies (U5 and U6) are directed to epitopes outside the binding domain, whereas the other two (T1 and T6) interact at or near the ligand-binding domain. ¹⁸ Purified nonimmune mouse IgGs were from ICN Biomedicals (Aurora, IL). 

Purification of IgG

Antibodies from rabbit antisera, nonimmune sera, or ascites fluid were purified on a protein A Sepharose column (Sigma, St. Louis, MO), as described. ¹⁹ Briefly, samples were added to columns equilibrated with phosphate-buffered saline (PBS; pH 8). Columns were washed extensively with PBS (pH 8) and IgG eluted with 0.1 M sodium acetate (pH 3) into tubes with 1 M Tris hydrochloride (pH 8) to neutralize the acid. The concentration of antibodies was determined by the Bradford method (Bio-Rad, Richmond, CA). 

Enzyme-linked Immunosorbent Assay

Polyclonal antibodies directed against 23-kDa PRL or 16-kDa PRL and the INN-rPRL-1 anti-23-kDa PRL monoclonal antibody were characterized by enzyme-linked immunosorbent assay (ELISA), performed as described. ²⁰ For these assays, wells were coated with 10 ng of either 23-kDa PRL or 16-kDa PRL. Primary polyclonal (1:2000) or monoclonal (1:100) antibodies, with or without serial dilutions of 23-kDa or 16-kDa PRLs were added to the coated wells. Horseradish peroxidase–conjugated anti-rabbit IgG or anti-mouse IgG (Vector Laboratories, Burlingame, CA) was used at 1:2000 or 1:1000 dilution, respectively. Bound horseradish peroxidase–conjugated antibodies were revealed by reaction with 2 mg of ortho-phenylenediamine (dihydrochloride) in 0.1 M citrate buffer (pH 5.0) with 0.1% hydrogen peroxide. Reaction was stopped with 4 N sulfuric acid and optical density (OD) determined at 490 nm. 

Western Blot Analysis

Twenty-three-kDa and 16-kDa PRLs were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in 15% acrylamide separating and 4% stacking laboratory gels, under reducing conditions using the buffer system of Laemmli. ²¹ Proteins were blotted into nitrocellulose membranes and probed with polyclonal (1 μg/ml) or monoclonal (2 μg/ml) antibodies. Antigen–antibody reaction was developed with the use of an alkaline phosphatase secondary antibody kit (Bio-Rad, Hercules, CA). 

Rat Corneal Bioassay

The bioassay was performed with male Wistar rats (200–250 g), as previously described. ²² Briefly, rats were anesthetized with sodium pentobarbital. A corneal pocket was made by inserting a 27-gauge needle, with the pocket’s base 1 mm from the limbus. Implants containing the test substances were made using poly-2-hydroxylethylmethacrylate (Hydron; Interferon Sciences, New Brunswick, NJ), stained with india ink, and combined (1:1) with the sample substance. The Hydron substance solution was pipetted onto a sterile Teflon surface and air dried to produce a 2-mm-diameter disc. Approximately one eighth of the disc was implanted in each corneal pocket. Animals were killed 6 days after implantation and corneas dissected, flattened by cuts leaving three to four regions of at least 4 mm² each, and fixed for 2 hours with 4% paraformaldehyde-1.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). Corneal capillary network was delineated by staining the enclosed red blood cells by a peroxidase reaction, as described. ¹⁰ Accordingly, visualization of vessels followed corneal incubation in 0.05% 3,3′-diaminobenzidine in the presence of 0.01% H₂O₂ in 0.1 M phosphate buffer. Subsequently, corneas were dehydrated, mounted (Entellan; Merck, Darmstadt, Germany), coverslipped, and analyzed in a light-field microscope (model 104; Nikon, Garden City, NY). Images were digitized with a video camera (model CCD72; Dage–MTI, Michigan City, IN) and subjected to analysis (NIH Image ver. 1.62, National Institutes of Health, Bethesda, MD). A mask of a region of interest, with a semitriangular shape and approximately 4.0 mm², was generated and centered at the implant on each corneal digitized image. This was the smaller common area to be contained within the segments delimited by the cuts performed to flattened the corneas. Corneal angiogenesis was quantitated by recording the OD associated with pixel gray values of stained capillaries within the region of interest mask. Image background was subtracted to ensure OD uniformity on the image. Only angiogenic reactions reaching an OD value equal to or higher than the double of the average control level were considered to be substantial enough for the evaluation of the incidence of neovascularization. In addition, the number of long vessels reaching the implant from perilimbal vessels was quantitated on the printed digitized corneal images by double-blind, independent observers. 

Immunocytochemistry

Male Wistar rats (200–250 g) were killed by decapitation. Eyes were enucleated and fixed for 2 hours with 4% paraformaldehyde-1.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4), transferred to 20% (wt/vol) sucrose until saturation, and then cut at 15 mm thickness in a cryostat (−16°C). Sections were incubated for 30 minutes in methanol containing 30% H₂O₂ and 1 hour in 10% bovine serum albumin (BSA), 3% normal calf serum, and 0.3% Triton X-100 in PBS. Subsequently, sections were incubated overnight in a humid atmosphere at 4°C with each anti-PRL receptor monoclonal antibody (100 μg/ml) in 0.3% Triton X-100 PBS. The sections were rinsed and incubated with secondary biotinylated anti-mouse IgG followed by avidin-biotin-peroxidase (ABC kit, Vector Laboratories). Bound anti-PRL receptor antibodies were revealed with 0.05% 3,3′-diaminobenzidine and 0.15% H₂O₂ for 2 to 3 minutes. Sections were dehydrated, mounted (Entellan; Merck), coverslipped, and analyzed in a light-field microscope (model BX60; Olympus, Lake Success, NY). 

Statistical Analysis

Data reported as ODs and number of vessels (mean ± SEM) were analyzed by two-way analysis of variance (replicate × treatment) to partition out variation due to experimental replicates. Mean values for each treatment were compared using Duncan’s new multiple-range test. Differences were considered significant at P < 0.05. Data reported as incidence of angiogenesis were analyzed for statistical significance by the nonparametric Mann–Whitney rank sum test. 

Results

Inhibition by 16-kDa PRL of Corneal Angiogenesis Induced by bFGF

bFGF stimulated corneal angiogenesis in a dose-dependent manner (Figs. 1 and 2) . Increasing amounts of bFGF (from 60 ng to 1.7 μg) were associated with increasing ODs in an area surrounding the implants, that were significantly higher than those observed with vehicle (Hydron alone) or BSA-containing implants (Fig. 2A) . Values in these two control groups were similar to those recorded in equivalent corneal areas distant from the implants (data not shown). Dose-related increases in the proportion of corneas with substantial angiogenic responses (OD more than two times the values of untreated corneas) were also observed (Fig. 2B) . The implantation of 60 and 120 ng bFGF elicited substantial angiogenesis in 1 of 7 and in 14 of 18 corneas, respectively, and higher doses resulted in substantial angiogenesis in all the treated corneas (Fig. 2B) . Implants containing BSA did not induce substantial angiogenesis in any of the 17 treated corneas. 

To investigate the putative antiangiogenic effect of 16-kDa and 23-kDa PRL, corneal angiogenesis was stimulated with 0.25 μg bFGF (the lowest dose observed to induce substantial angiogenesis in 100% of corneas) in the absence or presence of 2 μg of the two PRL standards (Fig. 3) . Inclusion of 16-kDa PRL in the implants reduced the magnitude of the angiogenic responses induced by bFGF by more than 65% (Figs. 3A 3B) . Likewise, the incidence of corneas with substantial angiogenic responses was reduced by half by treatment with 16-kDa PRL. Conversely, cotreatment with bFGF and 23-kDa PRL resulted in an incidence and magnitude (OD) of angiogenesis similar to those induced by bFGF alone (Fig. 3B) . 

Characterization of Anti-PRL Antibodies

Anti-23-kDa PRL polyclonal and monoclonal antibodies were characterized by ELISA and western blot analysis(Fig. 4) . Competition curves using different concentrations of 16-kDa or 23-kDa PRL standards showed that although both PRL standards reacted with the monoclonal antibody at similar potencies, 16-kDa PRL had a higher immunoreactivity than 23-kDa PRL for both anti-23-kDa and anti-16-kDa polyclonal antibodies (Fig. 4A) . 

Western blot analysis probed with all the PRL antibodies detected 23-kDa PRL and 16-kDa PRL standards with similar intensities (Fig. 4B) . No reaction was observed with either PRL standard when western blot analyses were probed with preimmune rabbit or mouse antibodies (data not shown). None of the antibodies reacted with the BSA nonspecific control. 

Induction of Rat Corneal Angiogenesis by Anti-PRL Antibodies

Implants containing anti-16-kDa PRL or anti-23-kDa PRL polyclonal antibodies induced corneal angiogenesis (Fig. 5) . Although anti-16-kDa PRL antibodies induced substantial angiogenic responses in 20% of the corneas treated, anti-23-kDa PRL antibodies did the same in 40% (Fig. 5) . On average, these responses were equivalent in magnitude to those induced by doses of bFGF between 60 and 120 ng (Fig. 2) . Although OD determinations adequately monitor the moderate and strong neovascularization responses induced by bFGF, they could underestimate mild vessel growth, such as that occurring in response to PRL antibodies. Thus the number of leading vessels reaching the implant was used as an additional score for proangiogenic actions of anti-PRL antibodies. Both anti-16-kDa PRL and anti-23-kDa PRL antibodies increased the number of these vessels over preimmune antibody values. To control for nonspecific reactions to antigen–antibody interactions, 36 corneas were treated with 2 μg BSA, anti-BSA antibodies, or their combination, with no positive responses detected (data not shown). 

Similar results were obtained with the use of the monoclonal anti-23-kDa PRL antibody (Fig. 6) . A 47% incidence of neovascularization, accompanied by a larger number of vessels, was observed in corneas implanted with anti-PRL monoclonal antibody (Fig. 6) . One third to one fifth of these responses, however, appeared to be nonspecific, because they were elicited by nonimmune mouse antibodies. 

PRL Receptor Immunoreactivity in Corneal Epithelium

The presence of PRL receptors in corneal epithelium was investigated in an attempt to elucidate a possible extraocular origin for PRL in the cornea. Here, PRL receptors could function as an avenue for the incorporation into the cornea of PRL known to be present in tears. ²³ The U5 anti-PRL receptor monoclonal antibody detected PRL-receptorlike immunoreactivity in epithelial cells throughout the cornea (Fig. 7) . Specificity of immunostaining was validated by a similar positive reaction observed with the other three anti-PRL-receptor monoclonal antibodies used (U6, T1, and T6; data not shown) and by the lack of a reaction in the absence of primary antibodies or after their replacement with nonimmune mouse IgGs (Fig. 7) . 

Discussion

Neovascularizations are the most frequent causes of blindness worldwide. Intraocular neovascularization occurs in, among other disorders, diabetic retinopathy, ischemic retinal vein occlusion, retinopathy of prematurity, corneal transplantation, and ocular trachoma. Angiogenesis in the eye may be the result of an imbalance between stimulatory and inhibitory factors that presumably occurs from the elevated expression of local angiogenic factors induced by ischemia. ^{24 25} Various angiogenic factors have been proposed to mediate vasoproliferative eye diseases, including bFGF, insulin-like growth factor-1, and, most important, vascular endothelial growth factor. ^{24 25 26 27} However, the imbalance responsible for the pathologic angiogenesis may also result from downregulation of inhibitors of blood vessel growth. ^{4 5 6}  

Several angiogenesis inhibitors have been described, some of which correspond to fragments of larger proteins. ⁵ The latter include angiostatin, a 36-kDa internal fragment of plasminogen ²⁸ ; endostatin, the 20-kDa C-terminal fragment of collagen XVIII ²⁹ ; an internal fragment of platelet factor 4 ³⁰ ; fragments of laminin ³¹ ; peptides derived from thrombospondin ³² ; and the 14-kDa and 16-kDa N-terminal fragments of PRL. ^{10 12}  

PRL fragments of 14 kDa and 16 kDa inhibit endothelial cell proliferation. ^{8 9 10} Similarly, 16-kDa PRL inhibits the activation of urokinase, ¹¹ a protease involved in endothelial cell migration, and tube formation by endothelial cells in vitro. ¹⁰ Moreover, 16-kDa PRL inhibits the in vivo formation of blood vessels in the chick embryo CAM. ¹⁰ PRL fragments with antiangiogenic effects appear to be produced by the enzymatic proteolysis of 23-kDa PRL in both the anterior and posterior lobes of the pituitary gland and are found in the circulation. ^{11 12 33} Moreover, endothelial cells from different blood vessels and species have been shown to express the PRL gene, ^{34 35 36} and endothelia-derived PRL molecules have been found to affect endothelial cell proliferation. ³⁴ The placenta produces proliferin and PRP, two PRL-like proteins that stimulate and inhibit angiogenesis, respectively. ¹³ Both proteins have been implicated in the regulation of placental neovascularization. ¹³ Accordingly, various lines of evidence provide support to the hypothesis that members of the PRL family function as potential regulators of angiogenesis. 

In the present study we used the rat cornea to investigate this hypothesis. The cornea provides a well characterized bioassay for angiogenesis, in which the elicitation of an angiogenic reaction is a convincing demonstration of true neovascularization, because this organ is normally avascular. ¹⁴ With this assay we have evaluated the in vivo antiangiogenic action of 16-kDa PRL on the well-known angiogenic effect of bFGF. In addition, we have treated corneas with anti-PRL antibodies in an attempt to unmask possible PRL-like proteins involved in tissue avascularity. Antibody administration to corneas has been a valuable tool for the detection of molecules active in angiogenesis. ³⁷  

Corneal angiogenesis was quantitated through measurement of densitometric (OD) values associated with stained capillaries, which also allowed a qualitative determination of the incidence of angiogenesis. These parameters provided reliable dose–response quantitations of bFGF-induced angiogenesis. Active doses of bFGF were similar to effective ones previously described in this assay. ³⁷ Addition of 16-kDa PRL reduced bFGF-induced angiogenesis by 65%, when judged by the ODs of the treated area. Moreover, 16-kDa PRL reduced by 50% the incidence of bFGF-induced angiogenesis. Sixteen-kDa PRL was active at a dose five times lower than the effective one in the CAM assay. ¹⁰ This in vivo inhibitory effect of 16-kDa PRL is consistent with the previously described antiangiogenic effect of PRP, ¹³ the placental PRL-like protein that competes for 16-kDa PRL receptor in endothelial cells. ⁹ PRP was shown to inhibit bFGF-induced vessel growth in rat corneas. ¹³ However, 23-kDa PRL did not modify bFGF angiogenic action. This absence of effect is to be expected because 23-kDa PRL does not bind to 16-kDa PRL receptors in endothelial cells. ⁹ Receptors mediating the antiangiogenic actions of PRL fragments differ from the classic (cloned) PRL receptors, both structurally and functionally. ⁹ Thus, 23-kDa PRL has been shown to be inactive in angiogenesis, in endothelial cells both in vitro and in vivo, and in the microvasculature of the CAM. ^{8 10}  

The in vivo antiangiogenic action of 16-kDa PRL was originally described in the CAM. ¹⁰ The CAM in the early chick embryo (days 6–9) is a rapidly growing membrane where angiogenesis actively occurs in response to multiple stimulatory factors. In the present study we showed that 16-kDa PRL inhibited the in vivo effect of a defined stimulator of angiogenesis, bFGF. Inhibition by 16-kDa PRL of the in vivo angiogenic effect of bFGF matches the reported 16-kDa PRL inhibition of bFGF-induced proliferation of endothelial cells in culture. ^{8 10} Actually, it is known that this PRL fragment inhibits the mitogenic effect of bFGF and vascular endothelial growth factor by acting distal to their receptors and proximal to the mitogen-activated protein kinases, specifically by inhibition of the activation of Raf-1. ^{38 39} The fact that 16-kDa PRL inhibits the in vivo effect of bFGF is consistent with the notion that PRL-like proteins can counteract the net stimulatory effect of angiogenic factors. 

In an attempt to investigate whether endogenous PRLs are involved in the control of cornea avascularity, we studied the effect of neutralizing the action of endogenous PRL with PRL-directed antibodies on corneal neovascularization. The polyclonal and monoclonal PRL antibodies used reacted with both 23-kDa and 16-kDa PRL standards by ELISA and western blot. All these antibodies resulted in corneal angiogenic responses. The number of long capillary vessels was increased in the area surrounding antibody-releasing implants. Accordingly, treatment with anti-PRL antibodies was associated with higher ODs in these areas, and therefore, with a substantial stimulation of the incidence of angiogenesis. Specificity of antibody action was supported by significant differences between nonimmune rabbit and mouse IgG. However, mouse IgG produced a moderate angiogenic reaction, perhaps because of inflammation. Nevertheless, the anti-PRL monoclonal antibody induced a significant response three times higher than this control. The possibility that inflammation related to antigen–antibody interaction contributed to the angiogenic effect of anti-PRL antibodies is unlikely, because no angiogenic response followed cotreatment of corneas with BSA and anti-BSA antibodies. 

The possibility that angiogenesis induced by PRL antibodies is caused by blockage of the antiangiogenic effect of endogenous PRL-like molecules, raises various intriguing issues. This novel proposal requires that PRL-like molecules, specifically PRL fragments and/or other antiangiogenic PRL forms, be present in the cornea. Although these PRLs must be determined and identified, their presence in the cornea is likely. We have recently detected immunoreactive PRL in the aqueous fluid of humans, ⁴⁰ and rat retinal endothelial cells have been found to express the PRL gene. ³⁶ Moreover, PRL receptors have been identified in the retina, ⁴¹ and PRL has been proposed to play a role in the metamorphosis of visual pigments in amphibians. ⁴² In addition, PRL has been shown to be synthesized by acinar cells of lacrimal glands. ⁴³ These cells express the PRL mRNA and locate PRL-like immunoreactivity in secretory vesicles. ⁴³ Consistent with this, PRL has been detected in human tears, ²³ and evidence has been provided to suggest that its presence in tears is under systemic hormonal control. ⁴⁴ Actually, it has been proposed that this hormone may have a role in the physiology of the cornea, the conjunctiva, or the conjunctiva-associated lymphoid tissue. ⁴³ In this regard, it can be hypothesized that antiangiogenic PRLs present in the cornea could originate from lacrimal PRL and thus could have an extraocular origin. Consistent with this proposal, immunocytochemistry with four well-characterized PRL receptor monoclonal antibodies, showed that receptors for prolactin seemed to be located throughout the corneal epithelium. Because PRL is known to be taken up by receptor-mediated endocytosis in several cell types, ^{45 46} it is possible to hypothesize that epithelial cells in the cornea may bind and internalize the PRL present in tears, and thus provide an avenue for the incorporation of lacrimal PRL into this organ. Therefore, these results add support to the presence of PRL in the cornea and provide an insight into the possible extraocular origin of this hormone. 

No conclusions can be drawn on the nature of the possible contribution of PRLs to cornea avascularity. Sixteen-kDa PRL may be a candidate, because it inhibits corneal neovascularization, and it reacts with all the antibodies used. Alternatively, other PRL-like proteins may be involved, as suggested by the fact that the 16-kDa PRL-directed antibodies were less active in corneal blood vessel growth than antibodies raised against the whole 23-kDa PRL molecule. 

The present study provides observations consistent with the hypothesis that members of the PRL family may function as regulators of angiogenesis, specifically in the maintenance of corneal avascularity. The presence of PRL receptors in ocular tissues similar to those of the cornea further supports the eye as a novel target for PRL actions. Identification of PRL in ocular tissues and fluids and its possible ocular and/or extraocular origin are under current investigation. 

²  Permanent address: Fundación Universitaria de Boyacá, Tunja, Colombia.

 Supported by Grants 3662P,27950-N from the National Council of Science and Technology of Mexico, IN201895 from the National University of Mexico, 75197-554801 from The Howard Hughes Medical Institute, and RGA 96-097 from The Third World Academy of Sciences.

 Submitted for publication November 24, 1998; revised March 24, 1999; accepted May 20, 1999.

 Commercial relationships policy: N.

 

Figure 1.

 

bFGF-induced corneal angiogenesis. Representative corneas treated for 6 days with Hydron (Interferon Sciences) implants containing increasing doses of bFGF. Corneal capillaries were revealed by staining the enclosed red blood cells by peroxidase reaction. The region of interest used for evaluating angiogenic responses is delineated in the cornea containing no bFGF. Numbers below corneas indicate micrograms of bFGF contained in the respective implant.

Figure 2.

 

bFGF-induced corneal angiogenesis. (A) Angiogenic responses induced by increasing doses of bFGF, quantitated by the OD of a 4-mm² area surrounding the implant. Values represent the mean ± SEM of the number of corneas shown in parentheses. (B) Incidence of angiogenesis, defined as the proportion of corneas with substantial angiogenesis (ODs more than two times those of untreated corneas). Numbers above bars represent the number of corneas with substantial angiogenesis over the number of treated corneas.* P < 0.05 versus vehicle- or BSA-treated controls.

Figure 3.

 

Inhibition by 16-kDa PRL of bFGF-induced corneal angiogenesis. (A) Representative corneas implanted with bFGF (0.25 μg) alone or together with 16-kDa PRL (2 μg). (B) Angiogenic responses to bFGF (0.25 μg) in the absence or presence of 2 mg 16-kDa PRL or 23-kDa PRL, quantitated by the OD associated with stained capillaries. (C) Incidence of angiogenesis (as defined in legend to Fig. 2 ) after treatment of corneas. (B) Other conditions are as described in Figures 1 and 2 . *P < 0.05 versus vehicle-implanted control; **P < 0.05 versus group treated with bFGF alone.

Figure 4.

 

Characterization of anti-PRL antibodies. (A) ELISA competition curves using serial concentrations of 16-kDa (▵) or 23-kDa (•) PRL standards and anti-23-kDa PRL polyclonal antibodies (α23-kDa PRL), monoclonal antibodies (α23-kDa PRL Mab), or an anti-16-kDa PRL polyclonal antibody (α16-kDa PRL). (B) Twenty-three-kDa and 16-kDa rat PRL standards, revealed by western blot analysis probed with the respective anti-PRL polyclonal and monoclonal antibodies. None of the antibodies reacted with the BSA-treated negative control.

Figure 5.

 

Induction of corneal angiogenesis by anti-PRL polyclonal antibodies. (A) Representative corneas implanted with nonimmune antibodies from normal rabbit serum (rIgGs; 3 μg), anti-16-kDa PRL (a16-kDa PRL; 2 μg), or anti-23-kDa PRL (a23-kDa PRL; 2 μg) polyclonal antibodies. (B) Corneal angiogenesis in response to the respective antibodies was evaluated by quantitation of the OD associated with stained capillaries (left), the incidence of angiogenesis (middle), and the number of long vessels reaching the implants (right). Other conditions are as described in Figures 1 and 2 . *P < 0.05 versus treatment with nonimmune rabbit antibodies.

Figure 6.

 

Induction of corneal angiogenesis by an anti-PRL monoclonal antibody. (A) Representative corneas implanted with vehicle (control) or with 2 μg of either nonimmune mouse antibodies (mIgG) or anti-23-kDa PRL monoclonal antibody (a23-kDa PRL Mab). (B) Corneal angiogenesis in response to the respective antibodies was evaluated by the quantitation of the OD associated with stained capillaries (left), the incidence of angiogenesis (middle), and the number of vessels reaching the implants (right). Other conditions are as described in Figures 1 and 2 . *P < 0.05 versus vehicle-treated control.

Figure 7.

 

Immunoperoxidase identification of the PRL-receptor in corneal epithelial cells. Top: immunostaining with U-5 anti-PRL-receptor monoclonal antibody (100 μg/ml). Bottom: control section where nonimmune mouse IgG (100μ g/ml) was substituted for primary antibodies. Scale bar, 40 μm.

The authors thank Gabriel Nava and Leopoldo González-Santos for their expert technical assistance.