- Email: [email protected]
Immunohistopathologic evaluation of choroidal neovascular membranes following Verteporfin-photodynamic therapy
Immunohistopathologic Evaluation of Choroidal Neovascular Membranes Following Verteporfin-photodynamic Therapy SALVATORE GRISANTI, MD, OLCAY TATAR, MD, SERAP CANBEK, BART A. LAFAUT, MD, FAIK GELISKEN, MD, WERNER INHOFFEN, PHD, PETER SZURMAN, MD, SABIN AISENBREY, MD, JOLANTA OFICJALSKA-MLYNCZAK, MD, AND KARL ULRICH BARTZ-SCHMIDT, MD
● PURPOSE: To evaluate the vascularization and proliferative activity in choroidal neovascular membranes due to age-related macular degeneration after verteporfin photodynamic therapy and submacular removal. ● DESIGN: Interventional case series. ● METHODS: In a retrospective review of seven patients who underwent removal of subfoveal classic choroidal neovascular membranes after treatment with photodynamic therapy 3 to 146 days earlier, membranes were stained for CD 34, CD 105, and Ki-67 and correlated with clinical pictures and fluorescein angiography. ● RESULTS: Fluorescein angiography performed on the day of surgery disclosed nonperfusion of the treated area 3 days after photodynamic therapy, but perfusion and leakage were seen at greater post–photodynamic therapy intervals. Membranes excised 3 days after photodynamic therapy showed CD34 and CD105 positive, mostly occluded vessels. The endothelial cells appeared damaged. Ki-67 activity was low. In membranes excised 34 to 146 days after photodynamic therapy, all vessels appeared patent and were lined by healthy endothelial cells with strong expression of CD34 and CD105. Ki-67 expression was elevated after 34 days but decreased thereafter. Biosketches and/or additional material at www.ajo.com Accepted for publication Dec 12, 2003. From the Department of Ophthalmology I (S.G., O.T., S.C., F.G., W.I., P.S., K.U.B.S.), University of Tuebingen, Tuebingen, Germany; Department of Ophthalmology (B.A.L.), AZ StJan, Bruges, Belgium; Department of Ophthalmology, University of Wroclaw, Poland (J.O-M.); and Department of Neuroscience (S.A.), Tufts University, Boston, Massachusetts, USA. This work was supported by the Grimmke Foundation, Jung Foundation, Tubitak Foundation and Vitreoret Foundation. Grisanti and Tatar contributed equally to this work. Inquires to: Assistant Professor Salvatore Grisanti, MD, Department of Ophthalmology I, Division of Vitreoretinal Surgery, Eberhard-Karls University of Tuebingen, Schleichstrasse 12–15, 72076 Tuebingen, Germany; fax: (⫹49) 7071-295215; e-mail: [email protected]
914
©
2004 BY
● CONCLUSION:
Photodynamic therapy did not cause a general or complete occlusion of vessels within the choroidal neovascular membranes, as suggested by fluorescein angiography 3 days postintervention, but the endothelial cells appeared to be severely damaged. Proliferative activity within these specimens was reduced. At longer intervals after photodynamic therapy, the fibrovascular tissue seemed to recover; perfusion, hyperfluorescence, and leakage of the choroidal neovascular membranes could be detected by fluorescein angiography. The clinical appearance showed a correlation with the immunohistologic characteristics of an increased proliferative activity and patent vascularization. (Am J Ophthalmol 2004;137:914 –923. © 2004 by Elsevier Inc. All rights reserved.)
A
GE-RELATED MACULAR DEGENERATION IS THE
leading cause of legal blindness in patients older than 65 years of age.1 Choroidal neovascular membranes are the main cause of visual loss.2 Photodynamic therapy with verteporfin (Visudyne, Novartis AG, Buelach, Switzerland) can delay visual loss and is accepted as a therapeutic option in subfoveal classic choroidal neovascular membranes.3 The desired mechanism of action of photodynamic therapy is a photothrombosis of the choroidal neovascularization, avoiding damage to overlying neurosensory retina.4 Evidence for this mechanism has been inferred by histopathologic studies in nonhuman primates and from angiographic evidence of nonperfusion of the treated area in clinical studies.5–7 Recent clinicopathologic studies have improved our understanding by evaluating the effect of photodynamic therapy on choroidal neovascular membranes by conventional histochemistry and electron microscopy.8 –11 We present our results of a clinical and immunohistologic study of surgically extracted choroidal neovascular membranes caused by age-related macular degeneration
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/04/$30.00 doi:10.1016/j.ajo.2003.12.049
TABLE 1. Clinical Characteristics
Case
Eye
Age/Sex
CNVM Type
Visual Acuity
PDT Treatments
Time to Surgery From the First PDT/Last PDT
1 2 3 4 5 6 7
L R L L L L L
76/M 78/F 54/M 84/M 83/M 81/M 78/F
Classic Classic Predom. classic Classic Classic Classic Classic
0.025 0.02 0.063 0.025 0.03 0.08 0.05
1 1 2 1 1 2 3
3 days 3 days 113/3 days 3 days 34 days 213/131 days 344/222/146 days
CNVM ⫽ choroidal neovascularization membrane; F ⫽ female; L ⫽ left; M ⫽ male; PDT ⫽ photodynamic therapy. R ⫽ right.
following photodynamic therapy. The analysis focuses on the vascularization and proliferative activity of the specimens extracted after different time intervals and various amounts of photodynamic therapy treatments. The level of vascularization and proliferative activity was determined by immunodetection of characteristic markers. CD34 was used as a panendothelial marker. CD105 (endoglin) has a pronounced expression in “activated” endothelial cells that are the constituent of vessels in tissues undergoing angiogenesis.12 Endoglin is an essential component of the TGF- receptor complex of human endothelial cells. It binds TGF-1 and TGF-3. TGF- is an angiogenic agent in vivo.13 Increased expression of endoglin by proliferating endothelial cells probably modulates their response to TGF- and hence regulates the angiogenic process.14 Finally, Ki-67 is a cell-cycle-associated antigen that is only present in proliferating cells, so it is used as a nuclear marker of cell proliferation.15
was below 20/200, which, according to the TAP-Investigation, is the minimum visual acuity required for recommending initial photodynamic therapy16,17 and 2) visual deterioration that had progressed following initial photodynamic therapy. Verteporfin photodynamic therapy was performed 3 days before macular surgery with the intention of reducing the bleeding of the lesion at the time of surgical extraction and the rate of recurrence. The experimental nature of the treatment procedure and the risks and benefits of all therapeutic options were discussed in detail. Each patient gave written informed consent after the nature of the procedure and alternatives had been fully explained. The study followed the guidelines of the declaration of Helsinki as revised in Tokyo and Venice and adhered to requirements of the local institutional review board. The histologic analysis of the specimens was approved by the institutional ethics committee. Clinical characteristics of the patients are summarized in Table 1.
METHODS
Tissue Preparation. Within minutes after surgery, excised choroidal neovascular membranes were fixed in 3.7% formalin and subsequently embedded in paraffin. Each specimen was serially sectioned into 5-m sections and mounted on poly-L-lysine coated glass slides (Dako, Glostrup, Denmark) for immunohistochemical staining. Hematoxylin-Eosin and Trichrome Masson staining were performed to determine the histologic orientation. Periodic acid-Schiffs (PAS) staining was used primarily to confirm the location of the diffuse drusen and to help the overall orientation of the specimens. Pertungsten acid hematoxylin was used to detect fibrin clots within the vessel lumina. The histologic appearance of the study specimens is described in detail in the results section.
WE RETROSPECTIVELY REVIEWED SEVEN EYES OF SEVEN
consecutive patients, in which submacular surgery for choroidal neovascular membranes was performed after verteporfin photodynamic therapy. All patients had a complete ophthalmologic examination including fundus photography and stereoscopic fluorescein angiography. Fluorescein angiography was performed before the verteporfin photodynamic therapy and thereafter on the day of surgery. According to the guidelines of the Macular Photocoagulation Study classification, all patients had subfoveal predominant classic choroidal neovascular membranes caused by age-related macular degeneration. The therapy options, including observation, conventional thermal laser photocoagulation, macular translocation with 360-degree retinotomy and choroidal neovascular membranes extraction were discussed with the patients. Surgical intervention and submacular removal of the choroidal neovascular membranes were offered when 1) visual acuity VOL. 137, NO. 5
Immunohistology. After serial paraffin sections were deparaffinized and rehydrated with a graded series of alcohol, various techniques for antigen retrieval were applied. For endoglin staining, antigen retrieval was accomplished by proteolytic digestion with 0.5% pronase (Sigma, Munich,
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
915
Germany). For Ki-67 staining, antigen retrieval was accomplished with citric buffer and heat treatment in a steamer. For CD34 staining, heat treatment for 2 minutes was sufficient. Immunohistochemical staining was performed according to the manufacturer’s protocol (Vectastain Universal Elite ABC PK-6200 kit, Vector Laboratories, Burlingham, California). To block endogenous peroxidase activities, 3% hydrogen peroxide and 0.1% sodium acid were applied to each section. Subsequently, the sections were incubated with horse serum (30 minutes). After several washing steps, the sections were incubated with the primary antibodies specific for human endoglin (Mouse, Mab, Clone SN6 hours, Dako), CD34 (Mouse, Mab, Immunotech, Hamburg, Germany), and Ki-67 (Mouse, Mab, Clone Ki-S5, Dako) for 12 hours at 4°C. Binding of the primary antibodies was detected with peroxidase-conjugated goat antimouse IgG (Dianova, Hamburg, Germany) after 1 hour of incubation at room temperature. After washing, antibody-treated sections were developed with 3-Diaminobenzidine (Fluka, Buchs, Germany), and subsequently counterstained with Mayer’s hematoxylin and coverslipping. For control specimens, the primary antibodies were substituted by appropriate normal sera. Vascularization was evaluated by analysis of the entire specimen with immunoperoxidase stain for CD34 and CD105 and counting the number of vessels. These numbers were then compared with the number of vessels recognizable in conventional histochemistry. Genuine proliferative activity was evaluated by counting the number of Ki-67 positive nuclei. The specimens were evaluated in a blinded fashion independently by two examiners.
disclosed a fibrovascular tissue. Fibrosis was nearly 90% (Table 2). Only a few capillaries and larger vascular structures were recognizable. Larger vessels with abnormal endothelial lining, with or without intraluminal debris, were identified. The intraluminal debris was identified as fibrin clots by pertungsten acid hematoxylin staining. The specimen contained retinal pigment epithelial cells and diffuse drusen. Identification of the vessels was made easier with immunohistologic staining (Figure 1, bottom left and right). The specimen probed with the panendothelial cell marker CD34 disclosed both patent and collapsed vessels (Figure 1, bottom left). Although severely damaged, most of the vessels were still expressing CD105, a marker for proliferating endothelial cells. The expression of this marker within the heavily damaged vessels suggested that these endothelial cells were angiogenically active before photodynamic therapy (Figure 1, bottom right). Ki-67, a nuclear marker for cell proliferation, was not expressed except for within a single cell. A reason for this may be that the expression of this nuclear antigen is more readily affected by photodynamic therapy than are membranebound antigens (Table 1). A similar clinical and histologic pattern could be also detected in cases 2, 3, and 4. Clinical characteristics are summarized in Table 1. In case 3, macular surgery was performed 3 days after a second photodynamic therapy. This case also presented similar histologic characteristics to those of cases 1, 2 and 4 (Table 2). Conventional histopathology demonstrated only a few recognizable capillaries and venules, giving an impression of a predominant fibrotic and minimally vascularized tissue. The patent vessels were lined by obviously damaged endothelial cells. Immunohistology, however, proved the existence of many more vessels, which although collapsed were identified by immunostaining. Staining for the proliferation marker Ki-67 antigen was completely negative in two cases. Only one positive cell could be detected in case 1, but nine cells stained positive for Ki-67 in case 4. Interestingly, the higher number of proliferating cells in case 4 was associated with a lower percentage of closed vessels.
Case 1. A 77-year-old man was referred with a 3 to 4 month history of blurred vision and visual acuity of 0.025 in his left eye. Ophthalmoscopy revealed a subretinal edema with macular hemorrhages (Figure 1, top left). A subfoveal classic choroidal neovascular membrane was seen on fluorescein angiography (Figure 1, top right). After detailed information about therapeutic options and visual prognosis, the patient opted to proceed with macular surgery. Additional photodynamic therapy 3 days before surgery was suggested to reduce the risk of intraoperative hemorrhage. The patient underwent treatment with verteporfin photodynamic therapy following standard guidelines. On the day of surgery, 3 days after photodynamic therapy, a hypofluorescence due to nonperfusion of the irradiated area and the choroidal neovascular membrane was seen in the early phases of angiography (Figure 1, middle left). Late-phase fluorescein angiography revealed hyperfluorescence and leakage at the fovea, consistent with choroidal ischemia (Figure 1, middle right). Macular surgery was performed without complication on the same day. Three months after macular surgery, visual acuity remained stable at 0.03, and there was no recurrent choroidal neovascular membrane. Conventional histology 916
AMERICAN JOURNAL
Case 5. An 83-year-old man presented with the complaint of decreased visual acuity in his left eye of a week’s duration. Visual acuity was 0.03 in the left eye. On ophthalmoscopy, there was subretinal edema and paramacular hemorrhage. On fluorescein angiography, subfoveal classic choroidal neovascular membranes were detected. One month after photodynamic therapy, there was still a hemorrhage on the border of the membrane (Figure 2, top left) and hyperflorescence due to leakage of the choroidal neovascular membranes, as visualized with fluorescein angiography (Figure 2, top right). Visual acuity was 0.025. The patient opted to proceed with submacular surgery. Choroidal neovascular membrane removal was performed 34 days after photodynamic therapy. Two weeks after surgery, visual acuity increased to 0.1. OF
OPHTHALMOLOGY
MAY 2004
FIGURE 1. Fundus photography and immunohistology of Case 1. (Top left) Red-free fundus photography and (top right) fluorescein angiography before photodynamic therapy. (Middle left) Early- and (middle right) late-phase fluorescein angiography 3 days (day of surgery) after photodynamic therapy. Immunohistochemistry of the extracted choroidal neovascular membrane probed with antibodies against CD34 (bottom left) and CD105 (bottom right), stained with 3-Diaminobenzidine resulting in a brown chromogen and counterstained with hematoxylin. The brown chromogen can be distinguished from the melanin granula (asterisks) contained in pigmented cells. The endothelial cell marker CD34 and CD105 are selectively expressed in vascular structures (e.g., arrows). Endothelial cells appear highly damaged (bottom left). Some vessels are collapsed and identified only by the positive staining (bottom right). Scale bars: 25 m (bottom left) and 50 m (bottom right).
VOL. 137, NO. 5
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
917
infiltration. The endothelial cells could be easily distinguished in most of the vessels. The expression of CD34 and CD105 was marked (Figure 3, bottom left and right). All vessels were patent. The proliferation marker Ki-67 could be detected in several cells but not as many as in case 5 (Table 2).
TABLE 2. Immunohistological Characteristics
Case
CD34*
CD105†
Ki-67‡
Fibrosis (%)
Occluded Vessel (%)
1 2 3 4 5 6 7
⫹ ⫹ ⫹⫹ ⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹
⫹ ⫹ ⫹⫹ ⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹
1 0 0 9 35 6 2
90 50 30 50 50 50 70
90 82 90 62.5 0 0 0
Case 7. A 78-year-old female patient presented with decreased visual acuity in her left eye of 0.16. A subfoveal classic choroidal neovascular membrane was diagnosed on fluorescein angiography. She underwent verteporfin photodynamic therapy. At follow-up, visual acuity decreased, and there was continued leakage on fluorescein angiography. Photodynamic therapy was repeated twice more at 3-month intervals. Two months after the third treatment with photodynamic therapy, visual acuity was 0.05, and there was a subretinal hemorrhage on the inferior border of membrane covering a large area. Five months later, a subfoveal classic choroidal neovascular membrane with a subretinal hemorrhage was seen (Figure 4, top left and right), and visual acuity was 0.067. After an extensive evaluation of alternative therapeutic possibilities, the patient opted to proceed with submacular surgery. The choroidal neovascular membrane was extracted, and visual acuity was 0.05 2 months after surgery; no recurrence of choroidal neovascular membranes was observed. The extracted specimen consisted of a prominent fibrotic area, confined by a more vascularized, rather cellular region with a lymphocytic cell infiltration. CD34 and CD105 were expressed by patent vessels (Figure 4, bottom left and right). Several Ki-67-positive cells could be detected, but these were far less in number than in case 5. Before surgery, fluorescein angiography showed less hyperflorescence than in cases 4 and 5.
*By analyzing of the specimen with immunoperoxidase stain for CD34 and counting the number of vessels: ⫹ ⫽ 1–15; ⫹⫹ ⫽ 16 –30; ⫹⫹⫹ ⫽ ⬎30. † By analysis of the specimen with immunoperoxidase stain for endoglin and counting the number of vessels: ⫹ ⫽ 1–15; ⫹⫹ ⫽ 16 –30; ⫹⫹⫹ ⫽ ⬎30. ‡ By analysis of the specimen with immunoperoxidase stain for Ki-67 and counting the number of positive nuclei.
Histopathologic examination revealed a very cellular, highly vascularized tissue with normal-appearing capillaries and larger vessels, often with prominent endothelial nuclei. Dense inflammatory infiltrate, lymphocytes and polymorphonuclear cells were seen. The endothelial cells had a strong expression of CD34 (Figure 2, middle left and right) and CD105 (Figure 2, bottom left), reflecting vital and active vessels. Many Ki-67 positive cells could be detected as well (Figure 2, bottom right). These proliferating cells mostly belonged to the inflammatory infiltrate, although some Ki-67 positive endothelial cells were also present. Case 6. An 81-year-old man presented with the complaint of visual decrease in his left eye for 2 weeks. Visual acuity in the left eye was 0.08. A subretinal edema with subretinal hemorrhage was seen by ophthalmoscopy. Classic subfoveal choroidal neovascularization was apparent on fluorescein angiography. The patient was treated with photodynamic therapy. Three months later, visual acuity was 0.06, and fluorescein angiography revealed leakage (Figure 3, top left). Verteporfin photodynamic therapy was repeated, and 4 months later visual acuity was 0.05. There was continued leakage, as indicated on fluorescein angiography (Figure 3, top right), and increased edema was observed on optical coherence tomography images. These findings and extensive discussion with the patient regarding treatment options led to the decision of macular surgery and choroidal neovascular membrane extraction. Six weeks later, visual acuity was 0.06, and no recurrent choroidal neovascular membrane were detected. Histology of the specimens disclosed a fibrovascular tissue with some areas of a rather dense lymphocytic cell 918
AMERICAN JOURNAL
DISCUSSION PHOTODYNAMIC THERAPY IS BASED ON A PHOTOCHEMI-
cal effect translated from a sensitizing agent to the exposed tissue. After stimulation by 689-nm laser, verteporfin is known to activate a singlet to an excited triplet state. This triplet state generates either reactive free radicals or singlet oxygen species. According to Schmidt-Erfurth and colleagues, endothelial damage from these free radicals causes the release of clotting factors and selective occlusion of neovascular membranes.4 Many questions about the mechanisms and prolonged effects of photodynamic therapy are still unanswered, however, and persistent leakage after photodynamic therapy is only one of these. In our series of seven patients, four membranes were extracted 3 days after photodynamic therapy. Fluorescein angiography on the day of surgery revealed a nonperfusion of the choroidal neovascular membranes and the laser spot area. This hypofluorescence was proposed as evidence of OF
OPHTHALMOLOGY
MAY 2004
FIGURE 2. Fundus photography and immunohistology of case 5. (Top left) Red-free fundus photography and (top right) fluorescein angiography 34 days after photodynamic therapy. Immunohistochemistry of the extracted choroidal neovascular membrane probed with antibodies against CD34 (middle left and right) and CD105 (bottom left). The brown chromogen demonstrates a highly vascularized specimen (middle left and right, bottom left). The vessels are patent and lined by a healthy-looking endothelium. Several cell nuclei express the proliferation marker Ki-67 (e.g., arrows). Scale bar: 100 m (middle left) and 50 m (middle and bottom rows).
VOL. 137, NO. 5
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
919
FIGURE 3. Fundus photography and immunohistology of case 6. (Top left) Fluorescein angiography before and (top right) 131 days after a second photodynamic therapy treatment. Immunohistochemistry of the extracted choroidal neovascular membrane probed with antibodies against CD34 (bottom left) and CD105 (bottom right) are stained with the brown chromogen. The specimen contains many patent vessels. Scale bars: 50 m (bottom left) and 100 m (bottom right).
choroidal neovascular membrane occlusion.4 Recent clinicopathologic reports, however, have demonstrated that choroidal neovascular membranes, 3 days after photodynamic therapy, contained some occluded vessels, but most were still patent.10,11 The histologic examination of the four specimens extracted 3 days after verteporfin photodynamic therapy in our study revealed damaged but patent vessels that were not obstructed by thrombi. Some of the patent vessels contained unclotted erythrocytes suggesting perfusion. Although fluorescein angiography suggested a nonperfusion of the choroidal neovascular membrane, the histologic results showed that this is not the case. Because conventional histology may not identify all occluded vessels, the reflected ratio of occluded vs nonoccluded vessels may be misleading. In fact, in our study the immunohistologic examination of the choroidal neovascular membranes revealed that the amount of occluded 920
AMERICAN JOURNAL
vessels was much higher than suggested by conventional histology. Therefore, a minimal amount of vessels with persisting perfusion may not always be prominent on fluorescein angiography. A further explanation for the prominent hypofluorescent treatment spot is given by Flower,18 who hypothesized that a reduced blood flow of the choriocapillaris affects the perfusion of the choroidal neovascular membranes. This hypothesis is supported by a recent study in photodynamic-therapy–treated eyes showing a substantial reduction of the choriocapillary perfusion by confocal indocyanine green angiography.19 Because most of the vessels in our choroidal neovascular membranes appeared occluded but some were still patent, we must assume that both the alteration of the main part of the choroidal neovascularization as well as the reduced perfusion of the choriocapillaris are responsible for the picture given by fluorescein angiography. OF
OPHTHALMOLOGY
MAY 2004
FIGURE 4. Fundus photography and immunohistology of Case 7. (Top left) Red free fundus photography and (top right) fluorescein angiography 146 days after a third photodynamic therapy treatment. Immunohistochemistry of the extracted choroidal neovascular membranes probed with antibodies against CD34 (bottom left) and CD105 (bottom right) shows several patent vessels stained by the brown chromogen. Scale bars: 50 m (bottom left) and 100 m (bottom right).
Nevertheless, the histologic findings were different when the choroidal neovascular membranes were removed later than 30 days following verteporfin photodynamic therapy. After this time interval, even by immunohistology, no occluded vessels could be detected. In contrast, the numerous patent vessels in choroidal neovascular membrane seemed strong and the lining endothelium healthy. This was also the case after two (case 6) and three (case 7) photodynamic therapy treatments. One question associated with this observation is whether the patent vessels are the result of regeneration and reperfusion of previously occluded vessels9 or whether they originate from neoangiogenesis after the treatment.8 Our findings favored the neoangiogenetic mechanism. The reason for this is based on the following: first, our study and other reports revealed the fact that in choroidal neovascular membranes exVOL. 137, NO. 5
tracted 3 days after photodynamic therapy, most if not all of the endothelial cells in both patent and occluded vessels look disintegrated and swollen.9 –11 The expression of the endothelial cell marker CD34 was notably reduced compared with specimens extracted after a longer interval. It is highly doubtful that these cells can regenerate, even if they were in an activated state, as shown by persisting CD105 expression. Second, the prominent expression of CD105 associated with the increased number of Ki-67-positive cells suggests strong proliferative activity within the choroidal neovascular membranes. Third, the vasoproliferative activity within the specimen is associated with the inflammatory infiltrate described in other studies.11 These characteristics are features of a wound-healing process. Although the cause is more complex, the forma-
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
921
tion of choroidal neovascular membranes due to agerelated macular degeneration is associated to some degree with wound-healing mechanisms. Additionally, any traumatizing event will enhance the wound-healing cascade with the initial characteristics of neoangiogenesis within an inflammatory setting. Schnurrbusch8 and Moshfeghi11 reported that new vessel formation is not present until a minimum of 63 days after such an event. In our study, however, strong proliferation activity and endoglin expression were observed in the choroidal neovascular membranes extracted only 34 days after photodynamic therapy. Verteporfin photodynamic therapy may be a mild and selective treatment, but it has still a traumatizing effect. The aim of photodynamic therapy is to occlude the vessels within the choroidal neovascular membranes, and this seems to be achieved in most vessels. Even patent vessels that can be detected 3 days after photodynamic therapy seem to be affected by the treatment. The question of whether these vessels remain patent or are reperfused within 3 days cannot yet be answered properly, but it is conceivable that these patent vessels are not the result of neoangiogenesis after the treatment. The disintegration of the pertinent endothelial cells does not support this idea. There is no question that the treated tissue is affected by photodynamic therapy, but this does not mean it has been destroyed. Therefore, the damaged tissue will induce a wound-healing process that includes an inflammatory response and the proliferation of endothelial and other cells. This is reflected in the immunohistologic results of our cases 5, 6, and 7. Additionally, in three patients (cases 2, 6, and 7) who underwent multiple photodynamic therapy, we found that repeated treatments did not affect the chronology of trauma followed by a reproliferative response. To the best of our knowledge, this is the first study to report immunohistologic changes regarding proliferative activity and endoglin staining in addition to histopathology and clinical findings in choroidal neovascular membranes previously treated by photodynamic therapy. The proper interpretation of this study, however, is limited by the small number of examined specimens and the fact that our cases may represent a negative selection. Although histopathologic findings in patients who profit from verteporfin photodynamic therapy may differ from our patients, it is conceivable that photodynamic therapy interferes with the process of scar formation, but the wound-healing process will start over, leading to a fibrotic, inactive scar. This theory is supported by the clinical outcome that verteporfin photodynamic therapy delays the progression of visual deterioration, but it definitively does not stop the process of neoangiogenesis and scar formation. The introduction of an adjunctive antiangiogenetic therapy to photodynamic therapy may therefore be a favorable approach to enhance the therapeutic success. 922
AMERICAN JOURNAL
REFERENCES 1. Bird AC, Bressler NM, Bressler SB, et al. An international classification and grading system for age-related maculopathy and age-related macular degeneration. The International ARM Epidemiological Study Group. Surv Ophthalmol 1995; 39:367–374. 2. Bressler NM, Bressler SB, Fine SL. Age-related macular degeneration. Surv Ophthalmol 1988;32:375–413. 3. Bressler NM, Bressler SB. Photodynamic therapy with verteporfin (Visudyne): Impact on ophthalmology and visual sciences. Invest Ophthalmol Vis Sci 2000;41:624 –628. 4. Schmidt-Erfurth U, Hasan T. Mechanisms of action of photodynamic therapy with verteporfin for the treatment of age related macular degeneration. Surv Ophthalmol 2000; 45:195–214. 5. Miller JW, Schmidt-Erfurth U, Sickenberg M, et al. Photodynamic therapy with verteporfin for choroidal neovascularization caused by age-related macular degeneration: Results of a single treatment in a phase 1 and 2 study. Arch Ophthalmol 1999;117:1161–1173. 6. Schmidt-Erfurth U, Miller JW, Sickenberg M, et al. Photodynamic therapy with verteporfin for choroidal neovascularization caused by age-related macular degeneration: Results of retreatments in a phase 1 and 2 study. Arch Ophthalmol 1999;117:1177–1187. 7. Reinke MH, Canakis C, Husain D, et al. Verteporfin photodynamic therapy retreatment of normal retina and choroid in the cynomolgus monkey. Ophthalmology 1999; 106:1915–1923. 8. Schnurrbusch UEK, Welt K, Horn L-C, Wiedemann P, Wolf S. Histological findings of surgically excised choroidal neovascular membranes after photodynamic therapy. Br J Ophthalmol 2001;85:1086 –1091. 9. Ghazi NG, Jabbour NM, de la Cruz ZC, Green WR. Clinicopathologic studies of age-related macular degeneration with classic subfoveal choroidal neovascularization treated with photodynamic therapy. Retina 2001;21:478 – 486. 10. Gelisken F, Lafaut B, Inhoffen W, Voelker M, Grisanti S, Bartz-Schmidt KU. Clinicopathological findings of choroidal neovascularisation following verteporfin photodynamic therapy. Br J Ophthalmol 2004;88:207–211. 11. Moshfeghi DM, Kaiser PK, Grossniklaus HE, et al. Clinicopathologic study after submacular removal of choroidal neovascular membranes treated with verteporfin ocular photodynamic therapy. Am J Ophthalmol 2003;135:343–350. 12. Kumar S, Ghellal A, Li C, et al. Breast carcinoma: Vascular density determined using CD 105 antibody correlates with tumor prognosis. Cancer Res 1999;59:856 –861. 13. Enenstein J, Waleh NS, Kramer RH. Basic FGF and TGF- differentially modulate integrin expression of human microvascular endothelial cells. Exp Cell Res 1992;203:499 –503. 14. Cheifetz S, Bellon T, Cales C, et al. Endoglin is a component of the transforming growth factor- receptor system in human endothelial cells. J Biol Chem 1922;267:19027– 19030. 15. Karak AK, Sarkar C, Chumber S, Tandon N. MIB-1 proliferative index in parathyroid adenoma and hyperplasia. Indian J Med Res 1997;105:235–238. 16. Treatment of age-related macular degeneration with Photodynamic therapy (TAP) study group. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin. One-year results of two randomized clinical trials-TAP Report 1. Arch Ophthalmol 1999;117:1329 –1345. OF
OPHTHALMOLOGY
MAY 2004
17. Treatment of age-related macular degeneration with Photodynamic therapy (TAP) study group. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin. Two-year results of 2 randomized clinical trials-TAP Report 2. Arch Ophthalmol 2001;119:198 –207. 18. Flower RW. Experimental studies of indocyanine green
VOL. 137, NO. 5
dye-enhanced photocoagulation of choroidal neovascularization feeder vessels. Am J Ophthalmol 2000;129:501– 512. 19. Schmidt-Erfurth U, Michels S, Barbazetto I, Laqua H. Photodynamic effects on choroidal neovascularization and physiological choroids. Invest Ophthalmol Vis Sci 2002;43: 830 –841.
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
923
● PURPOSE: To evaluate the vascularization and proliferative activity in choroidal neovascular membranes due to age-related macular degeneration after verteporfin photodynamic therapy and submacular removal. ● DESIGN: Interventional case series. ● METHODS: In a retrospective review of seven patients who underwent removal of subfoveal classic choroidal neovascular membranes after treatment with photodynamic therapy 3 to 146 days earlier, membranes were stained for CD 34, CD 105, and Ki-67 and correlated with clinical pictures and fluorescein angiography. ● RESULTS: Fluorescein angiography performed on the day of surgery disclosed nonperfusion of the treated area 3 days after photodynamic therapy, but perfusion and leakage were seen at greater post–photodynamic therapy intervals. Membranes excised 3 days after photodynamic therapy showed CD34 and CD105 positive, mostly occluded vessels. The endothelial cells appeared damaged. Ki-67 activity was low. In membranes excised 34 to 146 days after photodynamic therapy, all vessels appeared patent and were lined by healthy endothelial cells with strong expression of CD34 and CD105. Ki-67 expression was elevated after 34 days but decreased thereafter. Biosketches and/or additional material at www.ajo.com Accepted for publication Dec 12, 2003. From the Department of Ophthalmology I (S.G., O.T., S.C., F.G., W.I., P.S., K.U.B.S.), University of Tuebingen, Tuebingen, Germany; Department of Ophthalmology (B.A.L.), AZ StJan, Bruges, Belgium; Department of Ophthalmology, University of Wroclaw, Poland (J.O-M.); and Department of Neuroscience (S.A.), Tufts University, Boston, Massachusetts, USA. This work was supported by the Grimmke Foundation, Jung Foundation, Tubitak Foundation and Vitreoret Foundation. Grisanti and Tatar contributed equally to this work. Inquires to: Assistant Professor Salvatore Grisanti, MD, Department of Ophthalmology I, Division of Vitreoretinal Surgery, Eberhard-Karls University of Tuebingen, Schleichstrasse 12–15, 72076 Tuebingen, Germany; fax: (⫹49) 7071-295215; e-mail: [email protected]
914
©
2004 BY
● CONCLUSION:
Photodynamic therapy did not cause a general or complete occlusion of vessels within the choroidal neovascular membranes, as suggested by fluorescein angiography 3 days postintervention, but the endothelial cells appeared to be severely damaged. Proliferative activity within these specimens was reduced. At longer intervals after photodynamic therapy, the fibrovascular tissue seemed to recover; perfusion, hyperfluorescence, and leakage of the choroidal neovascular membranes could be detected by fluorescein angiography. The clinical appearance showed a correlation with the immunohistologic characteristics of an increased proliferative activity and patent vascularization. (Am J Ophthalmol 2004;137:914 –923. © 2004 by Elsevier Inc. All rights reserved.)
A
GE-RELATED MACULAR DEGENERATION IS THE
leading cause of legal blindness in patients older than 65 years of age.1 Choroidal neovascular membranes are the main cause of visual loss.2 Photodynamic therapy with verteporfin (Visudyne, Novartis AG, Buelach, Switzerland) can delay visual loss and is accepted as a therapeutic option in subfoveal classic choroidal neovascular membranes.3 The desired mechanism of action of photodynamic therapy is a photothrombosis of the choroidal neovascularization, avoiding damage to overlying neurosensory retina.4 Evidence for this mechanism has been inferred by histopathologic studies in nonhuman primates and from angiographic evidence of nonperfusion of the treated area in clinical studies.5–7 Recent clinicopathologic studies have improved our understanding by evaluating the effect of photodynamic therapy on choroidal neovascular membranes by conventional histochemistry and electron microscopy.8 –11 We present our results of a clinical and immunohistologic study of surgically extracted choroidal neovascular membranes caused by age-related macular degeneration
ELSEVIER INC. ALL
RIGHTS RESERVED.
0002-9394/04/$30.00 doi:10.1016/j.ajo.2003.12.049
TABLE 1. Clinical Characteristics
Case
Eye
Age/Sex
CNVM Type
Visual Acuity
PDT Treatments
Time to Surgery From the First PDT/Last PDT
1 2 3 4 5 6 7
L R L L L L L
76/M 78/F 54/M 84/M 83/M 81/M 78/F
Classic Classic Predom. classic Classic Classic Classic Classic
0.025 0.02 0.063 0.025 0.03 0.08 0.05
1 1 2 1 1 2 3
3 days 3 days 113/3 days 3 days 34 days 213/131 days 344/222/146 days
CNVM ⫽ choroidal neovascularization membrane; F ⫽ female; L ⫽ left; M ⫽ male; PDT ⫽ photodynamic therapy. R ⫽ right.
following photodynamic therapy. The analysis focuses on the vascularization and proliferative activity of the specimens extracted after different time intervals and various amounts of photodynamic therapy treatments. The level of vascularization and proliferative activity was determined by immunodetection of characteristic markers. CD34 was used as a panendothelial marker. CD105 (endoglin) has a pronounced expression in “activated” endothelial cells that are the constituent of vessels in tissues undergoing angiogenesis.12 Endoglin is an essential component of the TGF- receptor complex of human endothelial cells. It binds TGF-1 and TGF-3. TGF- is an angiogenic agent in vivo.13 Increased expression of endoglin by proliferating endothelial cells probably modulates their response to TGF- and hence regulates the angiogenic process.14 Finally, Ki-67 is a cell-cycle-associated antigen that is only present in proliferating cells, so it is used as a nuclear marker of cell proliferation.15
was below 20/200, which, according to the TAP-Investigation, is the minimum visual acuity required for recommending initial photodynamic therapy16,17 and 2) visual deterioration that had progressed following initial photodynamic therapy. Verteporfin photodynamic therapy was performed 3 days before macular surgery with the intention of reducing the bleeding of the lesion at the time of surgical extraction and the rate of recurrence. The experimental nature of the treatment procedure and the risks and benefits of all therapeutic options were discussed in detail. Each patient gave written informed consent after the nature of the procedure and alternatives had been fully explained. The study followed the guidelines of the declaration of Helsinki as revised in Tokyo and Venice and adhered to requirements of the local institutional review board. The histologic analysis of the specimens was approved by the institutional ethics committee. Clinical characteristics of the patients are summarized in Table 1.
METHODS
Tissue Preparation. Within minutes after surgery, excised choroidal neovascular membranes were fixed in 3.7% formalin and subsequently embedded in paraffin. Each specimen was serially sectioned into 5-m sections and mounted on poly-L-lysine coated glass slides (Dako, Glostrup, Denmark) for immunohistochemical staining. Hematoxylin-Eosin and Trichrome Masson staining were performed to determine the histologic orientation. Periodic acid-Schiffs (PAS) staining was used primarily to confirm the location of the diffuse drusen and to help the overall orientation of the specimens. Pertungsten acid hematoxylin was used to detect fibrin clots within the vessel lumina. The histologic appearance of the study specimens is described in detail in the results section.
WE RETROSPECTIVELY REVIEWED SEVEN EYES OF SEVEN
consecutive patients, in which submacular surgery for choroidal neovascular membranes was performed after verteporfin photodynamic therapy. All patients had a complete ophthalmologic examination including fundus photography and stereoscopic fluorescein angiography. Fluorescein angiography was performed before the verteporfin photodynamic therapy and thereafter on the day of surgery. According to the guidelines of the Macular Photocoagulation Study classification, all patients had subfoveal predominant classic choroidal neovascular membranes caused by age-related macular degeneration. The therapy options, including observation, conventional thermal laser photocoagulation, macular translocation with 360-degree retinotomy and choroidal neovascular membranes extraction were discussed with the patients. Surgical intervention and submacular removal of the choroidal neovascular membranes were offered when 1) visual acuity VOL. 137, NO. 5
Immunohistology. After serial paraffin sections were deparaffinized and rehydrated with a graded series of alcohol, various techniques for antigen retrieval were applied. For endoglin staining, antigen retrieval was accomplished by proteolytic digestion with 0.5% pronase (Sigma, Munich,
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
915
Germany). For Ki-67 staining, antigen retrieval was accomplished with citric buffer and heat treatment in a steamer. For CD34 staining, heat treatment for 2 minutes was sufficient. Immunohistochemical staining was performed according to the manufacturer’s protocol (Vectastain Universal Elite ABC PK-6200 kit, Vector Laboratories, Burlingham, California). To block endogenous peroxidase activities, 3% hydrogen peroxide and 0.1% sodium acid were applied to each section. Subsequently, the sections were incubated with horse serum (30 minutes). After several washing steps, the sections were incubated with the primary antibodies specific for human endoglin (Mouse, Mab, Clone SN6 hours, Dako), CD34 (Mouse, Mab, Immunotech, Hamburg, Germany), and Ki-67 (Mouse, Mab, Clone Ki-S5, Dako) for 12 hours at 4°C. Binding of the primary antibodies was detected with peroxidase-conjugated goat antimouse IgG (Dianova, Hamburg, Germany) after 1 hour of incubation at room temperature. After washing, antibody-treated sections were developed with 3-Diaminobenzidine (Fluka, Buchs, Germany), and subsequently counterstained with Mayer’s hematoxylin and coverslipping. For control specimens, the primary antibodies were substituted by appropriate normal sera. Vascularization was evaluated by analysis of the entire specimen with immunoperoxidase stain for CD34 and CD105 and counting the number of vessels. These numbers were then compared with the number of vessels recognizable in conventional histochemistry. Genuine proliferative activity was evaluated by counting the number of Ki-67 positive nuclei. The specimens were evaluated in a blinded fashion independently by two examiners.
disclosed a fibrovascular tissue. Fibrosis was nearly 90% (Table 2). Only a few capillaries and larger vascular structures were recognizable. Larger vessels with abnormal endothelial lining, with or without intraluminal debris, were identified. The intraluminal debris was identified as fibrin clots by pertungsten acid hematoxylin staining. The specimen contained retinal pigment epithelial cells and diffuse drusen. Identification of the vessels was made easier with immunohistologic staining (Figure 1, bottom left and right). The specimen probed with the panendothelial cell marker CD34 disclosed both patent and collapsed vessels (Figure 1, bottom left). Although severely damaged, most of the vessels were still expressing CD105, a marker for proliferating endothelial cells. The expression of this marker within the heavily damaged vessels suggested that these endothelial cells were angiogenically active before photodynamic therapy (Figure 1, bottom right). Ki-67, a nuclear marker for cell proliferation, was not expressed except for within a single cell. A reason for this may be that the expression of this nuclear antigen is more readily affected by photodynamic therapy than are membranebound antigens (Table 1). A similar clinical and histologic pattern could be also detected in cases 2, 3, and 4. Clinical characteristics are summarized in Table 1. In case 3, macular surgery was performed 3 days after a second photodynamic therapy. This case also presented similar histologic characteristics to those of cases 1, 2 and 4 (Table 2). Conventional histopathology demonstrated only a few recognizable capillaries and venules, giving an impression of a predominant fibrotic and minimally vascularized tissue. The patent vessels were lined by obviously damaged endothelial cells. Immunohistology, however, proved the existence of many more vessels, which although collapsed were identified by immunostaining. Staining for the proliferation marker Ki-67 antigen was completely negative in two cases. Only one positive cell could be detected in case 1, but nine cells stained positive for Ki-67 in case 4. Interestingly, the higher number of proliferating cells in case 4 was associated with a lower percentage of closed vessels.
Case 1. A 77-year-old man was referred with a 3 to 4 month history of blurred vision and visual acuity of 0.025 in his left eye. Ophthalmoscopy revealed a subretinal edema with macular hemorrhages (Figure 1, top left). A subfoveal classic choroidal neovascular membrane was seen on fluorescein angiography (Figure 1, top right). After detailed information about therapeutic options and visual prognosis, the patient opted to proceed with macular surgery. Additional photodynamic therapy 3 days before surgery was suggested to reduce the risk of intraoperative hemorrhage. The patient underwent treatment with verteporfin photodynamic therapy following standard guidelines. On the day of surgery, 3 days after photodynamic therapy, a hypofluorescence due to nonperfusion of the irradiated area and the choroidal neovascular membrane was seen in the early phases of angiography (Figure 1, middle left). Late-phase fluorescein angiography revealed hyperfluorescence and leakage at the fovea, consistent with choroidal ischemia (Figure 1, middle right). Macular surgery was performed without complication on the same day. Three months after macular surgery, visual acuity remained stable at 0.03, and there was no recurrent choroidal neovascular membrane. Conventional histology 916
AMERICAN JOURNAL
Case 5. An 83-year-old man presented with the complaint of decreased visual acuity in his left eye of a week’s duration. Visual acuity was 0.03 in the left eye. On ophthalmoscopy, there was subretinal edema and paramacular hemorrhage. On fluorescein angiography, subfoveal classic choroidal neovascular membranes were detected. One month after photodynamic therapy, there was still a hemorrhage on the border of the membrane (Figure 2, top left) and hyperflorescence due to leakage of the choroidal neovascular membranes, as visualized with fluorescein angiography (Figure 2, top right). Visual acuity was 0.025. The patient opted to proceed with submacular surgery. Choroidal neovascular membrane removal was performed 34 days after photodynamic therapy. Two weeks after surgery, visual acuity increased to 0.1. OF
OPHTHALMOLOGY
MAY 2004
FIGURE 1. Fundus photography and immunohistology of Case 1. (Top left) Red-free fundus photography and (top right) fluorescein angiography before photodynamic therapy. (Middle left) Early- and (middle right) late-phase fluorescein angiography 3 days (day of surgery) after photodynamic therapy. Immunohistochemistry of the extracted choroidal neovascular membrane probed with antibodies against CD34 (bottom left) and CD105 (bottom right), stained with 3-Diaminobenzidine resulting in a brown chromogen and counterstained with hematoxylin. The brown chromogen can be distinguished from the melanin granula (asterisks) contained in pigmented cells. The endothelial cell marker CD34 and CD105 are selectively expressed in vascular structures (e.g., arrows). Endothelial cells appear highly damaged (bottom left). Some vessels are collapsed and identified only by the positive staining (bottom right). Scale bars: 25 m (bottom left) and 50 m (bottom right).
VOL. 137, NO. 5
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
917
infiltration. The endothelial cells could be easily distinguished in most of the vessels. The expression of CD34 and CD105 was marked (Figure 3, bottom left and right). All vessels were patent. The proliferation marker Ki-67 could be detected in several cells but not as many as in case 5 (Table 2).
TABLE 2. Immunohistological Characteristics
Case
CD34*
CD105†
Ki-67‡
Fibrosis (%)
Occluded Vessel (%)
1 2 3 4 5 6 7
⫹ ⫹ ⫹⫹ ⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹
⫹ ⫹ ⫹⫹ ⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹
1 0 0 9 35 6 2
90 50 30 50 50 50 70
90 82 90 62.5 0 0 0
Case 7. A 78-year-old female patient presented with decreased visual acuity in her left eye of 0.16. A subfoveal classic choroidal neovascular membrane was diagnosed on fluorescein angiography. She underwent verteporfin photodynamic therapy. At follow-up, visual acuity decreased, and there was continued leakage on fluorescein angiography. Photodynamic therapy was repeated twice more at 3-month intervals. Two months after the third treatment with photodynamic therapy, visual acuity was 0.05, and there was a subretinal hemorrhage on the inferior border of membrane covering a large area. Five months later, a subfoveal classic choroidal neovascular membrane with a subretinal hemorrhage was seen (Figure 4, top left and right), and visual acuity was 0.067. After an extensive evaluation of alternative therapeutic possibilities, the patient opted to proceed with submacular surgery. The choroidal neovascular membrane was extracted, and visual acuity was 0.05 2 months after surgery; no recurrence of choroidal neovascular membranes was observed. The extracted specimen consisted of a prominent fibrotic area, confined by a more vascularized, rather cellular region with a lymphocytic cell infiltration. CD34 and CD105 were expressed by patent vessels (Figure 4, bottom left and right). Several Ki-67-positive cells could be detected, but these were far less in number than in case 5. Before surgery, fluorescein angiography showed less hyperflorescence than in cases 4 and 5.
*By analyzing of the specimen with immunoperoxidase stain for CD34 and counting the number of vessels: ⫹ ⫽ 1–15; ⫹⫹ ⫽ 16 –30; ⫹⫹⫹ ⫽ ⬎30. † By analysis of the specimen with immunoperoxidase stain for endoglin and counting the number of vessels: ⫹ ⫽ 1–15; ⫹⫹ ⫽ 16 –30; ⫹⫹⫹ ⫽ ⬎30. ‡ By analysis of the specimen with immunoperoxidase stain for Ki-67 and counting the number of positive nuclei.
Histopathologic examination revealed a very cellular, highly vascularized tissue with normal-appearing capillaries and larger vessels, often with prominent endothelial nuclei. Dense inflammatory infiltrate, lymphocytes and polymorphonuclear cells were seen. The endothelial cells had a strong expression of CD34 (Figure 2, middle left and right) and CD105 (Figure 2, bottom left), reflecting vital and active vessels. Many Ki-67 positive cells could be detected as well (Figure 2, bottom right). These proliferating cells mostly belonged to the inflammatory infiltrate, although some Ki-67 positive endothelial cells were also present. Case 6. An 81-year-old man presented with the complaint of visual decrease in his left eye for 2 weeks. Visual acuity in the left eye was 0.08. A subretinal edema with subretinal hemorrhage was seen by ophthalmoscopy. Classic subfoveal choroidal neovascularization was apparent on fluorescein angiography. The patient was treated with photodynamic therapy. Three months later, visual acuity was 0.06, and fluorescein angiography revealed leakage (Figure 3, top left). Verteporfin photodynamic therapy was repeated, and 4 months later visual acuity was 0.05. There was continued leakage, as indicated on fluorescein angiography (Figure 3, top right), and increased edema was observed on optical coherence tomography images. These findings and extensive discussion with the patient regarding treatment options led to the decision of macular surgery and choroidal neovascular membrane extraction. Six weeks later, visual acuity was 0.06, and no recurrent choroidal neovascular membrane were detected. Histology of the specimens disclosed a fibrovascular tissue with some areas of a rather dense lymphocytic cell 918
AMERICAN JOURNAL
DISCUSSION PHOTODYNAMIC THERAPY IS BASED ON A PHOTOCHEMI-
cal effect translated from a sensitizing agent to the exposed tissue. After stimulation by 689-nm laser, verteporfin is known to activate a singlet to an excited triplet state. This triplet state generates either reactive free radicals or singlet oxygen species. According to Schmidt-Erfurth and colleagues, endothelial damage from these free radicals causes the release of clotting factors and selective occlusion of neovascular membranes.4 Many questions about the mechanisms and prolonged effects of photodynamic therapy are still unanswered, however, and persistent leakage after photodynamic therapy is only one of these. In our series of seven patients, four membranes were extracted 3 days after photodynamic therapy. Fluorescein angiography on the day of surgery revealed a nonperfusion of the choroidal neovascular membranes and the laser spot area. This hypofluorescence was proposed as evidence of OF
OPHTHALMOLOGY
MAY 2004
FIGURE 2. Fundus photography and immunohistology of case 5. (Top left) Red-free fundus photography and (top right) fluorescein angiography 34 days after photodynamic therapy. Immunohistochemistry of the extracted choroidal neovascular membrane probed with antibodies against CD34 (middle left and right) and CD105 (bottom left). The brown chromogen demonstrates a highly vascularized specimen (middle left and right, bottom left). The vessels are patent and lined by a healthy-looking endothelium. Several cell nuclei express the proliferation marker Ki-67 (e.g., arrows). Scale bar: 100 m (middle left) and 50 m (middle and bottom rows).
VOL. 137, NO. 5
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
919
FIGURE 3. Fundus photography and immunohistology of case 6. (Top left) Fluorescein angiography before and (top right) 131 days after a second photodynamic therapy treatment. Immunohistochemistry of the extracted choroidal neovascular membrane probed with antibodies against CD34 (bottom left) and CD105 (bottom right) are stained with the brown chromogen. The specimen contains many patent vessels. Scale bars: 50 m (bottom left) and 100 m (bottom right).
choroidal neovascular membrane occlusion.4 Recent clinicopathologic reports, however, have demonstrated that choroidal neovascular membranes, 3 days after photodynamic therapy, contained some occluded vessels, but most were still patent.10,11 The histologic examination of the four specimens extracted 3 days after verteporfin photodynamic therapy in our study revealed damaged but patent vessels that were not obstructed by thrombi. Some of the patent vessels contained unclotted erythrocytes suggesting perfusion. Although fluorescein angiography suggested a nonperfusion of the choroidal neovascular membrane, the histologic results showed that this is not the case. Because conventional histology may not identify all occluded vessels, the reflected ratio of occluded vs nonoccluded vessels may be misleading. In fact, in our study the immunohistologic examination of the choroidal neovascular membranes revealed that the amount of occluded 920
AMERICAN JOURNAL
vessels was much higher than suggested by conventional histology. Therefore, a minimal amount of vessels with persisting perfusion may not always be prominent on fluorescein angiography. A further explanation for the prominent hypofluorescent treatment spot is given by Flower,18 who hypothesized that a reduced blood flow of the choriocapillaris affects the perfusion of the choroidal neovascular membranes. This hypothesis is supported by a recent study in photodynamic-therapy–treated eyes showing a substantial reduction of the choriocapillary perfusion by confocal indocyanine green angiography.19 Because most of the vessels in our choroidal neovascular membranes appeared occluded but some were still patent, we must assume that both the alteration of the main part of the choroidal neovascularization as well as the reduced perfusion of the choriocapillaris are responsible for the picture given by fluorescein angiography. OF
OPHTHALMOLOGY
MAY 2004
FIGURE 4. Fundus photography and immunohistology of Case 7. (Top left) Red free fundus photography and (top right) fluorescein angiography 146 days after a third photodynamic therapy treatment. Immunohistochemistry of the extracted choroidal neovascular membranes probed with antibodies against CD34 (bottom left) and CD105 (bottom right) shows several patent vessels stained by the brown chromogen. Scale bars: 50 m (bottom left) and 100 m (bottom right).
Nevertheless, the histologic findings were different when the choroidal neovascular membranes were removed later than 30 days following verteporfin photodynamic therapy. After this time interval, even by immunohistology, no occluded vessels could be detected. In contrast, the numerous patent vessels in choroidal neovascular membrane seemed strong and the lining endothelium healthy. This was also the case after two (case 6) and three (case 7) photodynamic therapy treatments. One question associated with this observation is whether the patent vessels are the result of regeneration and reperfusion of previously occluded vessels9 or whether they originate from neoangiogenesis after the treatment.8 Our findings favored the neoangiogenetic mechanism. The reason for this is based on the following: first, our study and other reports revealed the fact that in choroidal neovascular membranes exVOL. 137, NO. 5
tracted 3 days after photodynamic therapy, most if not all of the endothelial cells in both patent and occluded vessels look disintegrated and swollen.9 –11 The expression of the endothelial cell marker CD34 was notably reduced compared with specimens extracted after a longer interval. It is highly doubtful that these cells can regenerate, even if they were in an activated state, as shown by persisting CD105 expression. Second, the prominent expression of CD105 associated with the increased number of Ki-67-positive cells suggests strong proliferative activity within the choroidal neovascular membranes. Third, the vasoproliferative activity within the specimen is associated with the inflammatory infiltrate described in other studies.11 These characteristics are features of a wound-healing process. Although the cause is more complex, the forma-
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
921
tion of choroidal neovascular membranes due to agerelated macular degeneration is associated to some degree with wound-healing mechanisms. Additionally, any traumatizing event will enhance the wound-healing cascade with the initial characteristics of neoangiogenesis within an inflammatory setting. Schnurrbusch8 and Moshfeghi11 reported that new vessel formation is not present until a minimum of 63 days after such an event. In our study, however, strong proliferation activity and endoglin expression were observed in the choroidal neovascular membranes extracted only 34 days after photodynamic therapy. Verteporfin photodynamic therapy may be a mild and selective treatment, but it has still a traumatizing effect. The aim of photodynamic therapy is to occlude the vessels within the choroidal neovascular membranes, and this seems to be achieved in most vessels. Even patent vessels that can be detected 3 days after photodynamic therapy seem to be affected by the treatment. The question of whether these vessels remain patent or are reperfused within 3 days cannot yet be answered properly, but it is conceivable that these patent vessels are not the result of neoangiogenesis after the treatment. The disintegration of the pertinent endothelial cells does not support this idea. There is no question that the treated tissue is affected by photodynamic therapy, but this does not mean it has been destroyed. Therefore, the damaged tissue will induce a wound-healing process that includes an inflammatory response and the proliferation of endothelial and other cells. This is reflected in the immunohistologic results of our cases 5, 6, and 7. Additionally, in three patients (cases 2, 6, and 7) who underwent multiple photodynamic therapy, we found that repeated treatments did not affect the chronology of trauma followed by a reproliferative response. To the best of our knowledge, this is the first study to report immunohistologic changes regarding proliferative activity and endoglin staining in addition to histopathology and clinical findings in choroidal neovascular membranes previously treated by photodynamic therapy. The proper interpretation of this study, however, is limited by the small number of examined specimens and the fact that our cases may represent a negative selection. Although histopathologic findings in patients who profit from verteporfin photodynamic therapy may differ from our patients, it is conceivable that photodynamic therapy interferes with the process of scar formation, but the wound-healing process will start over, leading to a fibrotic, inactive scar. This theory is supported by the clinical outcome that verteporfin photodynamic therapy delays the progression of visual deterioration, but it definitively does not stop the process of neoangiogenesis and scar formation. The introduction of an adjunctive antiangiogenetic therapy to photodynamic therapy may therefore be a favorable approach to enhance the therapeutic success. 922
AMERICAN JOURNAL
REFERENCES 1. Bird AC, Bressler NM, Bressler SB, et al. An international classification and grading system for age-related maculopathy and age-related macular degeneration. The International ARM Epidemiological Study Group. Surv Ophthalmol 1995; 39:367–374. 2. Bressler NM, Bressler SB, Fine SL. Age-related macular degeneration. Surv Ophthalmol 1988;32:375–413. 3. Bressler NM, Bressler SB. Photodynamic therapy with verteporfin (Visudyne): Impact on ophthalmology and visual sciences. Invest Ophthalmol Vis Sci 2000;41:624 –628. 4. Schmidt-Erfurth U, Hasan T. Mechanisms of action of photodynamic therapy with verteporfin for the treatment of age related macular degeneration. Surv Ophthalmol 2000; 45:195–214. 5. Miller JW, Schmidt-Erfurth U, Sickenberg M, et al. Photodynamic therapy with verteporfin for choroidal neovascularization caused by age-related macular degeneration: Results of a single treatment in a phase 1 and 2 study. Arch Ophthalmol 1999;117:1161–1173. 6. Schmidt-Erfurth U, Miller JW, Sickenberg M, et al. Photodynamic therapy with verteporfin for choroidal neovascularization caused by age-related macular degeneration: Results of retreatments in a phase 1 and 2 study. Arch Ophthalmol 1999;117:1177–1187. 7. Reinke MH, Canakis C, Husain D, et al. Verteporfin photodynamic therapy retreatment of normal retina and choroid in the cynomolgus monkey. Ophthalmology 1999; 106:1915–1923. 8. Schnurrbusch UEK, Welt K, Horn L-C, Wiedemann P, Wolf S. Histological findings of surgically excised choroidal neovascular membranes after photodynamic therapy. Br J Ophthalmol 2001;85:1086 –1091. 9. Ghazi NG, Jabbour NM, de la Cruz ZC, Green WR. Clinicopathologic studies of age-related macular degeneration with classic subfoveal choroidal neovascularization treated with photodynamic therapy. Retina 2001;21:478 – 486. 10. Gelisken F, Lafaut B, Inhoffen W, Voelker M, Grisanti S, Bartz-Schmidt KU. Clinicopathological findings of choroidal neovascularisation following verteporfin photodynamic therapy. Br J Ophthalmol 2004;88:207–211. 11. Moshfeghi DM, Kaiser PK, Grossniklaus HE, et al. Clinicopathologic study after submacular removal of choroidal neovascular membranes treated with verteporfin ocular photodynamic therapy. Am J Ophthalmol 2003;135:343–350. 12. Kumar S, Ghellal A, Li C, et al. Breast carcinoma: Vascular density determined using CD 105 antibody correlates with tumor prognosis. Cancer Res 1999;59:856 –861. 13. Enenstein J, Waleh NS, Kramer RH. Basic FGF and TGF- differentially modulate integrin expression of human microvascular endothelial cells. Exp Cell Res 1992;203:499 –503. 14. Cheifetz S, Bellon T, Cales C, et al. Endoglin is a component of the transforming growth factor- receptor system in human endothelial cells. J Biol Chem 1922;267:19027– 19030. 15. Karak AK, Sarkar C, Chumber S, Tandon N. MIB-1 proliferative index in parathyroid adenoma and hyperplasia. Indian J Med Res 1997;105:235–238. 16. Treatment of age-related macular degeneration with Photodynamic therapy (TAP) study group. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin. One-year results of two randomized clinical trials-TAP Report 1. Arch Ophthalmol 1999;117:1329 –1345. OF
OPHTHALMOLOGY
MAY 2004
17. Treatment of age-related macular degeneration with Photodynamic therapy (TAP) study group. Photodynamic therapy of subfoveal choroidal neovascularization in age-related macular degeneration with verteporfin. Two-year results of 2 randomized clinical trials-TAP Report 2. Arch Ophthalmol 2001;119:198 –207. 18. Flower RW. Experimental studies of indocyanine green
VOL. 137, NO. 5
dye-enhanced photocoagulation of choroidal neovascularization feeder vessels. Am J Ophthalmol 2000;129:501– 512. 19. Schmidt-Erfurth U, Michels S, Barbazetto I, Laqua H. Photodynamic effects on choroidal neovascularization and physiological choroids. Invest Ophthalmol Vis Sci 2002;43: 830 –841.
CHOROIDAL NEOVASCULAR MEMBRANES FOLLOWING PHOTODYNAMIC THERAPY
923