Influence of Photodynamic Therapy on Expression of Vascular Endothelial Growth Factor (VEGF), VEGF Receptor 3, and Pigment Epithelium–Derived Factor

Abstract

purpose. To evaluate the impact of photodynamic therapy (PDT) on expression and distribution of vascular endothelial growth factor (VEGF), VEGF receptor (VEGFR)-3, and pigment epithelium–derived factor (PEDF).

methods. Eyes of patients scheduled for enucleation due to untreatable malignancy served as study eyes (n = 4), age-matched donor eyes were used as the control (n = 4). PDT using verteporfin with the recommended standard parameters was applied to intact areas of the perimacular region. Lesions were classified by ophthalmoscopy, fluorescein angiography (FA), and indocyanine green angiography (ICGA), as well as light and electron microscopic (LM/EM) histology. Immunolabeling using specific antibodies against VEGF, VEGFR-3, and PEDF was performed in PDT-treated areas, untreated collateral areas in study eyes, and untreated areas of control eyes. Specimens were fixed in 4% paraformaldehyde and 1% glutaraldehyde and embedded in paraffin. Four-micrometer-thick sections were stained using the peroxidase-labeled streptavidin–biotin method.

results. All PDT-treated areas demonstrated characteristic choroidal hypofluorescence by FA and ICGA. LM/EM histology revealed selective damage of choriocapillary endothelial cells. VEGF was expressed in the endothelial layer of choriocapillaries and focally within larger choroidal vessels in treated areas, but not in untreated areas. Sites with positive VEGF labeling also demonstrated upregulation of VEGFR-3. PEDF expression was localized to retinas in all eyes; however, PEDF staining of choroidal endothelial cells was specific for treated areas of study eyes.

conclusions. PDT using verteporfin induces a reproducible angiogenic response in elderly human eyes. VEGF, VEGFR-3, and PEDF expression is enhanced after PDT. Choroidal endothelial cells appear to be the primary site of angiogenic stimulation.

Photodynamic therapy (PDT) leads to selective cytotoxicity of vascular endothelial cells through production of oxidative radicals, which is limited to a confined localization and a precise time interval. These qualities predispose the method as an ideal treatment approach for choroidal neovascularization (CNV) allowing alteration of the neovascular structure without interference with extravascular neural layers. ¹ This concept was supported by a series of randomized, placebo-controlled trials that demonstrated maintenance of visual acuity in patients with predominantly classic CNV due to age-related macular degeneration (AMD). ^{2 3} The functional benefit correlated well with the anatomic response showing a significant reduction in CNV growth and progressive resolution of leakage. ^{2 3 4} Consequently, verteporfin therapy has been approved by the medical and legal authorities and is currently used as an established treatment in thousands of patients worldwide. 

However, the potential and success of the approach are considerably compromised: CNV’s persistence and recurrence require frequent retreatments. Despite an obvious early effect angiographically, suggesting disappearance of CNV, patients experience a mean visual loss of two ETDRs during the first 6 months after initiation of treatment. ^{2 3} Obviously, the purely mechanistic approach does not apply to the biological system, with a multifactorial pathogenesis of CNV disease and the unknown biological effects of the phototoxic intervention. Analysis of PDT-induced vascular changes by indocyanine green angiography (ICGA), for example, has revealed unexpected perfusion changes in the collateral choroid after each PDT application. ⁵ Choroidal hypoperfusion persists during all ICGA phases and is often irreversible during follow-up. ⁶ Histology of PDT’s effects in human eyes have identified an occlusion of the choriocapillary layer after PDT applied using the recommended standard parameters. ^{7 8} Continuous recurrence, changes in vascular permeability, and choroidal occlusion are either the result or the source of biological tissue reactions such as the release or inhibition of angiogenic factors. 

Vascular endothelial growth factor (VEGF) is a major stimulator of CNV development and growth. Expression of VEGF is present in CNV surgically excised from human AMD-affected eyes, ^{9 10 11} is expressed in experimental CNV, and can be used to induce CNV in animal models. ^{12 13} Recent studies have highlighted the importance of a balance between stimulatory effects, represented by VEGF, and inhibitory effects most likely transferred by pigment epithelium–derived factor (PEDF). ¹⁴ PEDF effectively inhibits the migration of endothelial cells in vitro and the in vivo development of experimental retinal neovascularization and CNV. ^{15 16 17} The impact of PDT on angiogenic mechanisms in the eye is completely unclear, because the angiogenic response to PDT has not been investigated experimentally or clinically. 

The purpose of this study was to evaluate the role of angiogenic stimulation or inhibition induced by the photodynamic process in human eyes. Standard verteporfin PDT was performed in retinochoroidal areas in eyes of patients scheduled for enucleation of an eye due to intraocular melanoma not amenable to conservative treatment. Treated sites were qualified as representative lesions based on characteristic choroidal hypofluorescence documented angiographically and thrombotic choriocapillary effects seen by light and electron microscopic (LM/EM) histology. Expression and distribution of VEGF protein (VEGF A), VEGF receptor type 3 (VEGF-3, or FLT-4) and PEDF protein were evaluated. 

Methods

Subjects

Eight eyes of eight donors were included in the study. Four eyes of patients who were undergoing enucleation for melanoma served as study eyes, and four donor eyes served as the control. The control eyes were from age-matched donors with no known ocular disease and were enucleated and fixed less than 5 hours after death. Death of the donors was caused by trauma, nonocular carcinoma, and heart failure. Patients in the study group had large malignant melanomas, too large for radiotherapy. Their mean age was 77.6 years (range, 72–87). The posterior pole was unaffected by the tumor mass as well as by the associated exudative retinal detachment. Scattered drusen were present in two eyes; the other two eyes showed no age-related changes. 

The study protocol was approved by the local ethics committee and adhered to the tenets of the Declaration of Helsinki regarding research involving human subjects. Informed consent was obtained from each patient with specific emphasis on the lack of any beneficial impact of treating the primary disease. 

Photodynamic Treatment and Lesion Characterization

PDT with verteporfin was performed according to the recommended standard procedure. ^{1 2 3 4} A solution of 6 mg/m² body surface verteporfin (Visudyne; Novartis, Bülach, Switzerland) was administered by an infusion of 10 minutes’ duration. Irradiation was started 5 minutes after the end of the infusion, using a conventional diode laser/slit lamp setting (Visulas; Zeiss, Jena, Germany). A single laser spot of 1.5 to 3 mm in diameter was applied to an intact chorioretinal area of the posterior pole, and the area was irradiated with a total exposure of 50 J/cm² over 83 seconds. 

Eyes were surgically removed 7 days after PDT. Clinical examination of the treated areas was performed before surgery on the same day and included ophthalmoscopy, fluorescein angiography (FA), and ICGA, using a scanning laser device (Heidelberg Retina Angiograph, Dossenheim, Germany). Images were taken in a confocal mode with a field of 30° during early-phase FA/ICGA, at 30 seconds to 1 minute, and during late phases at 10 minutes for FA and 20 minutes for ICGA. 

For histologic documentation of PDT-related choriocapillary damage the globes were immediately fixed in 4% paraformaldehyde and 1% glutaraldehyde in 0.1 M phosphate buffer for 5 days and processed for paraffin embedding. Lesion sites were localized based on angiographic images and bisected along the laser spots. For LM histology 8-μm-thick section were stained with hematoxylin and eosin and periodic acid Schiff (PAS). For electron microscopy, areas were postfixed in 2% buffered osmium tetroxide and embedded in epoxy resin (Epon). Semithin (1-μm-thick) sections were stained with toluidine blue. Ultrathin sections were treated with uranyl acetate-lead citrate and were examined with a transmission electron microscope (LEO 906E; Carl Zeiss Meditec, Oberkochen, Germany). 

Immunolabeling Procedure

Immunohistochemistry was performed on 4-μm-thick paraffin-embedded sections by the peroxidase-labeled streptavidin-biotin method, using a kit according to the manufacturer’s instructions (LSAB Plus; Dako, Glostrup, Denmark). A monoclonal mouse anti-human VEGF antibody (clone C-1; Santa Cruz Biotechnology, Santa Cruz, CA), an affinity-purified rabbit polyclonal antibodies against human VEGF receptor (VEGFR-3/FLT4; Chemicon International, Temecula, CA), and a monoclonal antibody against PEDF, raised against a synthetic peptide comprising amino acids 327-343 of human PEDF (courtesy of Petros Karakousis, F. M. Kirby Center for Molecular Ophthalmology, University of Pennsylvania, Philadelphia, PA) were used. ¹⁸  

Briefly, after proteolytic digestion with proteinase K (Dako) and blocking of endogenous peroxidase, the sections were successively incubated with the primary antibody dilutions (1:100), the biotinylated link antibody, and the peroxidase-conjugated streptavidin for 30 minutes each. 3-Amino 9-ethyl carbazole (AEC) was used as a chromogenic substrate, and Mayer’s hemalum was used as a counterstain. In negative control experiments, the primary antibody was omitted or replaced by equimolar concentrations of preimmune rabbit or mouse IgG or an irrelevant primary antibody. 

Immunolabeling for VEGF and VEGF receptor was performed in the PDT-treated areas and adjacent untreated areas, not exposed to photosensitization, from all four tumorous eyes and in corresponding areas from the posterior pole of four control eyes without prior drug or laser application. Antibodies against PEDF were used in PDT-treated areas and adjacent unexposed areas from the remaining two tumorous eyes and perimacular areas from the four control eyes. Immunoreactivity was analyzed separately in retinal structures, the RPE, and the choroidal layer. 

For a semiquantitative evaluation of staining intensity, a grading scheme was used indicating the degree of staining (Table 1) . Double plusses (++) indicate intense staining above the staining intensity observed in control eyes/areas; +, moderate labeling, comparable to background staining; (+), presence of weak expression of the antigen, below background, but present in areas showing no staining of these structures in control eyes; +/−, focal staining with an irregular pattern; and − absence of any staining. Intensity and presence of staining were determined by two masked observers (US-S, CC). 

Results

Angiographic and Histologic Characterization

Angiographic features were identical in all PDT lesions of all study eyes. By FA, lesions were shown to have mild transient hypofluorescence during early phases and were not distinguishable from background fluorescence during late phases. ICGA revealed hypofluorescence that persisted during the entire angiographic procedure. Lesions appeared as well-demarcated, homogeneously hypofluorescent spots during early sequences. The pattern of the larger choroidal vessels was visible within the treated spots, but choriocapillary filling was absent (Fig. 1A) . Leakage from the margins and vessels within the treatment site was observed during late-phase ICGA. 

LM histology showed consistent findings in all treated areas of the study eyes. Vascular endothelial cells of the choriocapillaries were selectively damaged with swelling, extensive vacuolization, shrinkage, and detachment of the basal lamina (Fig. 1B) . Capillary lumina were filled with deformed erythrocytes, degranulated platelets, and fibrin, whereas large vessels showed focal endothelial damage without thrombus formation. Examination by TEM confirmed primary endothelial damage with occlusion and degenerative changes of the choriocapillaris. Swelling, shrinkage, rupture of the plasma membrane, fragmentation, and complete decomposition of endothelial cells were abundantly present (Fig. 1C) . RPE cells of treated areas appeared intact, focally increased vacuolization was noted, but no destruction on the level of plasma membranes, cytoplasmic organelles, or nuclei or melanin granules were found. Retinal structures were unaffected, photoreceptors were structurally unchanged, and inner and outer segments were aligned regularly. Neural elements such as ganglion cells and the nuclear and plexiform layers were unremarkable. 

Expression of VEGF

Control eyes showed no immunoreactivity for VEGF. All retinal structures were completely negative (Fig. 2A) . No staining for VEGF was observed within the RPE or vascular endothelial cells. Areas in study eyes not exposed to sensitizing light showed no expression of VEGF at the level of the choriocapillaris or choroid. The retinas were generally negative except for some focal staining of individual ganglion cells in three of four eyes and of some astrocytes in one of four eyes (Fig. 2B) . Individual cells of the inner nuclear layer and foci of photoreceptor inner segments appeared to demonstrate mild reactivity. The RPE was completely negative in two of four eyes, weak staining of the RPE was seen in the remaining two eyes. 

PDT-treated areas of the study eyes exhibited a characteristic pattern of VEGF expression. An identical unspecific staining of focal cells in the ganglion cell layer was observed in the retinas (Fig. 2C) . Reactivity of inner photoreceptor segments was also documented, but labeling was more continuous and intensive than the focal appearance in the same areas in control eyes. No relevant staining exceeding that in the areas in control eyes was detected at the level of the RPE. A major difference in VEGF expression was obvious in the choroid. Intensive staining for VEGF was present within endothelial cells of the choriocapillaries, a consistent finding throughout all specimens with PDT-treated areas. A few scattered cells in the lumina of choroidal vessels and the surrounding extravascular stroma were also positive. In addition to specific labeling of the choriocapillary endothelium, focal reactivity for VEGF was present in the endothelia of deeper choroidal vessels at the site closest to the retina. Table 2 summarizes the reactivity of VEGF, VEGFR, and PEDF. 

Labeling of VEGFR-3

Staining for VEGFR-3 was generally less intense than that for VEGF protein, but the receptor was similar in expression and distribution. Control eyes exhibited no staining of any structure. Unexposed areas in control eyes and light-exposed areas in study eyes showed labeling of tumor-associated VEGF expression in the ganglion cell and inner nuclear layers. Two eyes in each group demonstrated staining at the level of the photoreceptors (Figs. 3A 3C) . As with VEGF labeling, a significant difference was found in VEGFR expression by components of the choroidal vasculature: Endothelial cells of the choriocapillaries revealed staining throughout the entire superficial capillary layer. Large choroidal vessels also demonstrated a positive, but more irregular, staining pattern, with focal areas of enhanced labeling (Figs. 3B 3D) . 

Immunoreactivity of PEDF

Retinal labeling was identical in control eyes and treated areas and untreated collateral areas of the study eyes. PEDF was observed in the retina and was restricted to the ganglion cell layer, the inner nuclear layer, and photoreceptor inner segments. The basal portion of the RPE was labeled in all specimens (Figs. 4A 4B) . However, although immunoreactivity was completely absent within vascular structures of the choroid in control eyes and untreated collateral areas in study eyes (Figs. 4C 4E) , uniform staining was observed throughout the choroidal endothelium of PDT-treated regions. Labeling of the choriocapillary band appeared more intense than reactivity in larger caliber vessels of the choroid (Figs. 4D 4F) . 

Discussion

PDT with verteporfin has become an established procedure predominantly to treat classic subfoveal CNV in AMD. Because a significant therapeutic benefit has been shown in solid clinical studies, treatment with PDT is indicated, and compromising effects such as frequent retreatments due to recurrence and choroidal malperfusion and particularly a reduced prognosis with respect to functional recovery appear to be acceptable. To improve the benefit of this potentially promising modality and reduce its limitations, a complete knowledge of the biological effects of PDT in the specific retinochoroidal setting is crucial. The purpose of this study was to identify the influence of PDT on the regulation of angiogenic factors such as VEGF and PEDF, which play an important role in the pathogenesis of neovascular AMD. To date, the interaction of PDT and angiogenic factors is completely unclear. Neither experimental in vivo nor in vitro studies have been performed, nor has there been any attempt to evaluate the impact of PDT on angiogenic processes in the human eye. 

PDT treatments using standard parameters were performed in a typical location, the retina of the posterior pole. Lesions were classified by angiography, which showed reproducible features consistent with choriocapillary hypofluorescence pathognomonic for PDT. Histology demonstrated classic thrombotic changes and uniformly identified the choriocapillary vascular endothelium as the primary site of structural damage. Immunolabeling revealed a clear influence of PDT on the expression of VEGF, VEGFR, and PEDF. VEGF immunoreactivity was upregulated in PDT-treated areas, whereas it was absent in adjacent unexposed regions and in the corresponding areas of control eyes. Specific reactivity with monoclonal antibodies identified the vascular endothelial cells of the choroid as the source of VEGF staining, which is most likely due to upregulated VEGF production rather than to VEGF binding. Accordingly, VEGFR-3 expression, a marker of a proliferative vascular endothelial phenotype, ¹⁹ was enhanced in PDT-treated areas and selectively localized at the level of the damaged choroidal endothelium. PEDF, usually a counterplayer of VEGF, showed a parallel reaction after PDT with an unexpected increase in expression that was also driven by vascular endothelial cells of the choroid. Upregulation of VEGF, VEGFR, and PEDF directly correlated with PDT-associated choriocapillary changes documented by angiography and histology. LM immunocytochemistry localized the antigen but is a semiquantitative approach. Qualitative localization was the major goal in this study, because VEGF, VEGFR-3, and PEDF staining was present only in the vascular endothelium of the choroid in treated areas and was completely absent in the choroid of adjacent unexposed areas in the same eye or in control eyes. 

A stimulating effect of PDT on release of VEGF is known in tumor therapy, although few investigators have quantified this phenomenon. ^{20 21} Ferrario et al. ²⁰ described substantially increased in vivo VEGF levels in a mouse mammary carcinoma after PDT and a lower increase in vitro in cultured tumor cells of the same species, suggesting a dual mechanism with a less intensive primary cellular response and a secondary more pronounced hypoxia-induced effect. PDT characteristically leads to damage of the microvasculature—namely, the endothelial cells and basement membrane—and establishes thrombogenic sits within the vascular lumina. A physiological cascade of thrombotic responses is initiated, including platelet aggregation, release of vasoactive molecules, leukocyte adhesion, increased vascular permeability, and vessel constriction. These effects result in hypoxia due to vascular obstruction and collapse, blood flow stasis, and tissue hemorrhages. ^{22 23} Depletion of tumor oxygenation during PDT leads to an increase in hypoxia markers. ²⁴ VEGF is a major stimulator for neovascularization due to ischemia. VEGF expression is upregulated by hypoxia. ^{25 26} Levels of VEGF are increased in the retina and vitreous of patients with ischemic retinopathies, ^{27 28} and VEGF alone is sufficient to induce neovascular growth, even in the absence of hypoxia. ^{29 30} Accordingly, VEGF receptor expression is influenced by hypoxic conditions, and VEGFR-3 is increased in diabetic eyes, mainly in leaky microvessels. ³⁰ Witmer et al. ³⁰ also showed that VEGF induces vascular expression of the VEGF-C/D receptor VEGFR-3. ³⁰ The results of our study suggest that a similar sequence of events occurs in ocular PDT. VEGFR-3, in contrast to VEGFR-1 and -2, in the adult is mainly restricted to lymphatic endothelium, ^{31 32} but becomes upregulated on proliferating blood vessels as well. ¹⁹  

PDT-related hypoxia is mainly induced by occlusion of choroidal vasculature. Choroidal occlusion after PDT was underestimated for a long time because choroidal effects are not easily detected by FA, which was the only established angiographic modality used in the PDT trials. ^{2 3 4} Analysis of ICGA images clearly demonstrated reproducible and persistent choroidal hypofluorescence. ^{5 6} Histology in human eyes showed that characteristic ICGA hypofluorescence was consistent with a thrombosis of choriocapillaries and individual larger choroidal vessels. ^{7 8} The observation of an intensive and dose-related damage to the choriocapillaris and choroid, in fact, triggered the present study focusing on the biological consequences of the structural damage associated with standard PDT. 

Histopathologic reports in the monkey model have shown that choroidal damage was present in experimental CNV using a fluence of 150 J/cm², whereas choroidal damage in human eyes was seen at light doses as low as 50 J/cm². ³³ A cumulative dose–response was seen angiographically and histologically with more severe damage to the retina and choroid at higher doses that eventually led to chronic absence of the choriocapillaris. ³⁴ Long-term follow-up has shown reperfusion of the choriocapillaris in normal choroid to occur at 7 weeks. ³⁵ In the rat model, damage to areas of normal choroid and retina varied as a function of the verteporfin and light doses. ³⁶ Accordingly, the relation of the dosage and the extent of choroidal damage suggest that the standard treatment used in human eyes represents an overdose effect. 

The primary site of damage was identified at the level of choroidal vascular endothelial cells, which showed swelling, ruptures of the plasma membrane, and fragmentation to complete degeneration (Fig. 1B) . Endothelial cells of the choroid are also the source of increased VEGF expression and VEGF receptor upregulation after the standard PDT procedure in human eyes as demonstrated by immunohistochemistry performed 1 week after PDT injury. A nonspecific binding of the antibodies to the PDT-treated area is unlikely, because only monoclonal antibodies against VEGF and PEDF were used, and a specific demarcation of vascular endothelial cells and not other cell types within the choroidal stroma was seen. All controls were consistently negative. Although our model does not contain an additional CNV structure, considering the dimension of a small CNV lesion compared with the large area of collateral choroidal damage suggests that the choroidal effect plays an important role in treated eyes with AMD-related CNV. PDT-induced choroidal lesions are particularly large, because a safety margin of 1000 μm is added to the largest linear diameter of the lesion. 

A clinicopathologic correlation in a patient treated with PDT has been reported by Greene et al. ³⁷ Evidence of endothelial cell degeneration with platelet aggregation and thrombus formation was noted. The same features were found in the choroid of PDT-treated areas, suggesting an identical sequence of events in CNV as in choroidal structures. 

Use of a model with the preexisting presence of VEGF in AMD eyes with CNV lesions would prevent the documentation of PDT-induced VEGF changes. Obviously, tumor-related angiogenic changes have to be considered in our model, but the VEGF response of the choroid was seen consistently and as a unique feature in PDT-treated areas only. However, the focal reactivity of cells in the ganglion cell layer, inner nuclear layer, and inner photoreceptor segments in all study eyes in general was attributed to the angiogenic stimulus of the intraocular malignancy. Vinores et al. ³⁸ demonstrated staining for VEGF intraretinally in 46% of subjects in the proximity of melanomas (i.e., retinal vessel walls, ganglion cells, inner or outer nuclear layers, and retinal pigment epithelium). VEGF expression in the choroid was not reported in melanoma-affected eyes and was not found in unexposed areas of tumorous eyes in this study. 

In contrast to VEGF, PEDF has an inhibitory effect on angiogenesis. ¹⁵ Expression of PEDF mRNA was detected in ganglion cells, cells of the inner nuclear layer, and RPE cells of normal rat eyes. ³⁹ Immunoreactivity for PEDF with identical localization was detected in the retinas of study and control eyes in our study highlighting the consistency of that finding in humans. PEDF is an antagonist of VEGF and inhibits VEGF-induced endothelial cell growth, migration, and the development of retinal neovascularization. ⁴⁰ In experimental CNV, PEDF, and PEDF mRNA were detected in proliferating RPE cells, macrophages, and fibroblasts within the photocoagulated lesion. ³⁹ PDT, a less invasive laser intervention, also induces PEDF expression. Not RPE, but endothelial cells of the choroidal vasculature appear to be the origin of PEDF. As we have shown, PEDF upregulation correlates anatomically with PDT-induced choriocapillary damage, in which hypoxia may be the promoting stimulus. This finding is consistent with the role of PEDF as a neuroprotective agent against ischemic retinal injury. ⁴¹ To date, PEDF expression after PDT has not been investigated. Presumably angiogenic inhibitors such as PEDF are responsible for the described reduction in human vascular cell migration after PDT. ⁴² Alternatively, recent experiments suggest that PEDF under certain conditions and in higher dosages can act synergistically with VEGF, inducing angiogenesis. ⁴³ Whether this aspect holds true for the situation after PDT remains to be studied. 

The novel observation of increased expression of VEGF, VEGFR, and PEDF explains a variety of clinical and angiographic features seen in PDT for neovascular AMD. The pronounced biological response highlights the invasive character of PDT, which leads to a mean loss of two lines in visual acuity during follow-up, despite anatomic and angiographic improvement. ^{2 3 4} Continuous recurrence during multiple PDT courses is probably the consequence of VEGF stimulation reinforced by the therapeutic intervention itself and closes the biological gap between angiographic evidence of choroidal occlusion with subsequent hypoxia and CNV regrowth. Choroidal malperfusion is most pronounced at 1 week after PDT, which correlates with enhanced VEGF release at the same time interval and most likely an increase in VEGFR-3, correlating with leaky microvessels. ^{30 44}  

Massive leakage from CNV and the surrounding choroid within days after PDT is consistent with the hyperpermeability typically induced by VEGF. According to a meta-analysis of Treatment of Age-Related Macular Degeneration with Photodynamic Therapy (TAP) and Verteporfin in Photodynamic Therapy (VIP) study data, lesion size appears to be the overwhelming factor predicting the prognosis of verteporfin therapy. ^{45 46} Less extensive choroidal hypoxia with less release of VEGF may explain the more favorable outcome of smaller lesions. The simultaneous release of PEDF may balance VEGF’s effects during follow-up. Under PEDF control, growth of CNV lesions may be reduced and leakage may subside progressively during long-term follow-up. Experimentally increased expression of PEDF substantially inhibits ocular neovascularization. ^{16 17}  

This study provides the first evidence for an angiogenic effect of PDT. Using immunolabeling techniques, increased expression of VEGF, VEGFR, and PEDF was shown in human eyes after verteporfin PDT. The presence and interaction of stimulatory and inhibitory agents explains contradictory clinical phenomena and has important implications for future treatment strategies. Obviously, combination therapy could be a promising option with the purpose of reducing PDT-associated VEGF release. PDT with adjunctive antiangiogenic treatment has improved tumoricidal responses experimentally. ^{19 47} Clinical trials combining PDT with anti-VEGF aptamers or angiostatic steroids are underway (Singerman L, IOVS 2002;43:ARVO E-Abstract 2908; Slakter J, et al. IOVS 2002;43:ARVO E-Abstract 2909). Mode, dosage, and timing of adjunctive drug application should be adapted to the pattern of PDT-induced angiogenic effects. Application of anti-VEGF compounds before PDT may block VEGF’s effects more efficiently. Diffusion of angiostatic steroids from the sub-Tenon space may be less efficient if the collateral choroid is already occluded and VEGF release is ongoing. Most important, the interference of PDT-induced angiogenic effects with the primary pathogenic mechanisms responsible for the development of neovascular AMD have to be analyzed. CNV formation and activity is the result of an imbalanced stimulatory and inhibitory condition that is further compromised by PDT. A decrease in choroidal perfusion and the resultant hypoxia are presumed promoters of CNV and are enhanced by PDT-induced release of VEGF. ^{48 49} The role of oxidative stress in the pathogenesis of neovascular AMD is well known. ⁵⁰ PDT invariably leads to the generation of free radicals and lipid peroxides that may further enhance expression of VEGF. ⁵¹ Continued preclinical and clinical studies are warranted to elucidate angiogenic effects, resolve PDT-related problems, and improve future combination strategies. 

 Submitted for publication November 1, 2002; revised March 5, 2003; accepted April 21, 2003.

 Disclosure: U. Schmidt-Erfurth (P), U. Schlötzer-Schrehard, None; C. Cursiefen, None; S. Michels, None; A. Beckendorf, None; G.O.H. Naumann, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Grading Scheme Used for Semiquantitative Evaluation of Staining Intensity

Degree of Staining	Code
Intense	++
Moderate	+
Weak	(+)
Focal	+/−
Absent	−

Figure 1.

 

Angiographic and histologic correlation. (A) PDT-treated areas demonstrated characteristic choroidal hypofluorescence within the area exposed to verteporfin and laser light. (B) LM histology of hypofluorescent areas revealed selective endothelial damage with subsequent thrombosis of the choriocapillary layer. Deeper choroidal vessels, RPE, and photoreceptor layer appeared intact (4-μm semithin section, toluidine blue). Magnification, ×250. (C) TEM histology of treated areas shows choriocapillaries obstructed by detached endothelial cells (En), platelets (Tr), fibrin (Fi), and red blood cells. Bar, 3 μm.

Figure 2.

 

Expression of VEGF immunolabeling for VEGF of retinal and choroidal structures of normal control eyes (A) as well as PDT-treated and untreated areas of study eyes 1 (B–E) and 2 (F, G). (A) VEGF staining was absent in retinas, RPE, and choroid of normal control eyes. (B) Unexposed area in study eye 1 revealed focal VEGF reactivity in retinal structures (ganglion cell layer, inner nuclear layer, and photoreceptor inner segments); choroidal structures, however, remained completely negative. (C) PDT-treated area in study eye 1 showed a similar but more pronounced staining pattern of retinal layers. Labeling of VEGF in choriocapillary endothelia was observed in treated regions only. (D) Study eye 1, untreated area shows no evidence of VEGF expression associated with choroidal structures at higher magnification. (E) Study eye 1, treated area shows VEGF protein abundantly expressed by choroidal vascular endothelia, particularly at the capillary level. (F) Study eye 2, untreated area: shows absence of VEGF in unexposed areas. (G) Study eye 2, treated area shows marked expression of VEGF visible along vascular endothelia of choriocapillaries. Magnification: (A–C) ×100; (D–G) ×250.

Table 2.

 

The Presence and Distribution of VEGF, VEGFR-3, and PEDF within the Retina, RPE and Choroid

Immunolabeling	Control					Study PDT-Treated					Study Unexposed
		Retina	RPE	Choroid	Retina	RPE	Choroid	Retina	RPE	Choroid
VEGF	−	−	−	+	(+)	++	+/−	(−)	−
				Focally in ganglion cell layer, inner nuclear layer, photoreceptor inner segments	Negative in 2/4	Choriocapillaries, retinal portion of large choroidal vessels	Focally in ganglion cell layer, inner nuclear layer, photoreceptor inner segments	Negative in 2/4; weak in 2/4
VEGFR-3	−	−	−	(+)	−	+	(+)	−	−
				Focally in ganglion cell layer, inner nuclear layer, photoreceptors		Choriocapillaries and focal areas of large choroidal vessels	Focally in ganglion cell and inner nuclear layers, and photoreceptors
PEDF	+		−	+		++	+		−
	Ganglion cell layer, inner nuclear layer, photoreceptor inner segments	Basal portion		Ganglion cell layer, inner nuclear layer, photoreceptor inner segments		Choriocapillaries, more intensively than in large choroidal vessels	Ganglion cell layer, inner nuclear layer, photoreceptor, inner segments

Figure 3.

 

Expression of VEGFR-3. Immunolabeling for VEGFR-3 of retinal and choroidal structures in PDT-treated and untreated areas of study eyes 1 (A, B) and 2 (C, D). (A) Study eye 1, untreated area shows negative VEGF receptor labeling within endothelia of the untreated choroid. (B) Study eye 1, treated area shows upregulated VEGF receptor expression in PDT-treated areas, associated with vascular endothelial cells of the choroid. (C) Study eye 2, untreated area shows absence of staining in the choroidal vasculature. (D) Study eye 2, treated area shows focal VEGFR expression in larger choroidal vessels, in addition to focal labeling of the choriocapillary walls. Magnification, ×250.

Figure 4.

 

Expression of PEDF immunolabeling in retinal and choroidal structures in PDT-treated and untreated areas of normal control eyes (A, B) and study eye 1 (C, D). (A) Control eyes demonstrate PEDF expression in the retina at the level of ganglion cells, the inner nuclear layer, and photoreceptor inner segments. (B) Higher magnification shows PEDF also expressed by the basal portions of the RPE cell layer (arrows). (C) Study eye 1, untreated area shows PEDF staining to be absent in the choroid of regions not exposed to PDT. (D) Study eye 1, treated area shows PDT-induced expression of PEDF by choroidal vascular endothelial cells. Magnification: (A) ×100; (B–D) ×250.