Selective Photodynamic Therapy by Targeted Verteporfin Delivery Toexperimental Choroidal Neovascularization Mediated by a Homing

Home , Verteporfin

LABORATORY SCIENCES Selective Photodynamic Therapy by Targeted Verteporfin Delivery to Experimental Choroidal Neovascularization Mediated by a Homing Peptide to Vascular Endothelial Growth Factor Receptor-2

Reem Z. Renno, MD; Yoshiko Terada, MD; Makhluf J. Haddadin, PhD; Norman A. Michaud, MS; Evangelos S. Gragoudas, MD; Joan W. Miller, MD

Objective: To evaluate the feasibility, efficacy, and se- ter treatment. Photodynamic therapy using verteporfin- lectivity of photodynamic therapy (PDT) using targeted PVA at the same drug dose achieved closure in 18 of 20 delivery of verteporfin to choroidal neovascularization CNV. Histological examination after PDT of normal retina (CNV) in the rat laser-injury model of CNV. and choroid using targeted verteporfin and irradiation at 1 hour showed minimal effect on retinal pigment epithe- Methods: We performed PDT in rat eyes on experi- lium and no injury to photoreceptors, whereas PDT using mental CNV and normal retina and choroid using verte- verteporfin-PVA resulted in retinal pigment epithelium ne- porfin conjugates. A targeted verteporfin conjugate was crosis and mild damage to photoreceptors. made by conjugating verteporfin (after isolation from its liposomal formulation) to a modified polyvinyl alcohol Conclusions: Verteporfin bound to the targeting pep- (PVA) polymer (verteporfin-PVA) followed by linkage tide, ATWLPPR, retained its spectral and photosensitiz- to the peptide ATWLPPR known to bind the receptor for ing properties. Angiography demonstrated localization vascular endothelial growth factor, VEGFR2. The verte- of the targeted verteporfin 1 hour after injection. Pho- porfin-PVA conjugate served as a control. We per- todynamic therapy using targeted verteporfin and the con- formed fluorescent fundus angiography to determine the trol conjugate were more effective in causing CNV clo- optimal timing of light application for PDT using the con- sure than standard liposomal verteporfin. The targeted jugates. Closure of CNV was assessed angiographically verteporfin resulted in more selective treatment than the and graded in a masked standardized fashion. We used control conjugate or standard verteporfin. These results standardized histological grading to compare the effects suggest that targeted PDT strategies based on selective on normal retina and choroid. expression of receptors on CNV vasculature may im- prove current therapy. Results: The verteporfin-PVA conjugation ratio was on average 28:1. The conjugate retained typical emission/ Clinical Relevance: Targeted PDT for CNV is feasible excitation spectra and photosensitizing activity and was as and may offer a qualitative improvement in current treat- efficient as an equivalent amount of verteporfin. Peak in- ments for patients with age-related macular degenera- tensity of targeted verteporfin in CNV was detected angio- tion. This study provides the basis for further preclini- graphically at 1 hour after intravenous injection. Photody- cal studies of targeted PDT strategies and subsequent namic therapy using targeted verteporfin (3 or 4.5 mg/m2) clinical trials. with light application 1 hour after drug injection showed angiographic closure of all treated CNV (17/17) 1 day af- Arch Ophthalmol. 2004;122:1002-1011

HOTODYNAMIC THERAPY and retinal vessels at the time that light is (PDT) using verteporfin as a applied. Preclinical work on PDT with photosensitizer has been verteporfin (QLT PhotoTherapeutics Inc, demonstrated in large clini- Vancouver, British Columbia) has shown cal trials to be an effective treatment-related damage to the surround- new treatment for subfoveal choroidal neo- ing retina, choroid, and retinal pigment P 5-7 vascularization (CNV) secondary to age- epithelium (RPE). This collateral dam- related macular degeneration and other age is cumulative with repeated PDT.7,8 causes.1-4 The preferential occlusion of Modifications to PDT, including combi- CNV after PDT as currently practiced is nation with antiangiogenic therapy9,10 or Author affiliations are listed based on the differences in biodistribu- with targeted photosensitizer, may im- at the end of this article. tion of the photosensitizer between CNV prove selectivity and vision outcomes.

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1002

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 Homing peptides are an emerging class of pharma- optical density, 250ϫ4.6 mm; a solvent system consisting of ceuticals that exploits differences between cell types by solution A (500 mL each of 1% [NH4]2SO4 and CH3CN and binding specific cell membrane receptors.11-13 The pep- 50 mL of CH3COOH) and solution B (500 mL of 1% [NH4]2 tide targeting of photosensitizers might enable specific SO4 and C4H8O and 50 mL of CH3COOH 50mL); a flow rate of and enhanced retention of photosensitizer to CNV and 1.7 mL/min; and a gradient of solutions A-B of 60%:40% for 5 minutes, then a starting gradient flow from 40% solution B to allow more selective PDT with minimal adverse effects. 70% solution B in 20 minutes and staying at 70% solution B Vascular endothelial growth factor (VEGF) expres- for 5 minutes before returning to 40% solution B. Verteporfin- sion and binding of VEGF to its kinase domain receptor PVA eluted at 8 to 10 minutes, whereas unconjugated verte- (KDR/FLK1 or VEGFR-2) is an important mediator of porfin eluted at 18 to 19 minutes. The molecular weight of verte- angiogenesis, including retinal and choroidal neovascu- porfin-PVA was determined by mass spectrometry and found larization.14,15 Inhibition of VEGF and VEGFR-2 pre- to be approximately 28 kDa (verteporfin-PVA ratio, approxi- vents retinal and choroidal neovascularization.16-18 The mately 28:1). presence of KDR or VEGFR-2 has been demonstrated in Before binding to the homing peptide, thiol groups were normal vessels but shows increased expression in endo- introduced to verteporfin-PVA using 3-mercaptopropionic acid thelial cells of neovascular tissue and is thus a potential (Sigma-Aldrich Corp) in acetate buffer (pH, 5.5). Coupling of candidate for peptide-mediated targeting of CNV.19,20 The verteporfin-PVA to the targeting peptide (ATWLPPR; molecu- peptide ATWLPPR is reported to specifically bind lar weight, 840 Da synthesized to our specifications by Anaspec, San Diego, Calif) was performed using sulfo-m-maleimidoben- VEGFR-2 and completely inhibit binding of native VEGF, 21 zoyl-N-hydroxysulfosuccinimide ester as a heterobifunctional thereby preventing VEGF-induced angiogenesis in vivo. cross-linking reagent in carbonate buffer (pH, 8.5). Products We propose to use ATWLPPR as a homing peptide were separated by means of high-performance liquid chroma- bound to verteporfin to target verteporfin to CNV by bind- tography with the solvent system of solution A, 100% water, ing to VEGFR-2 on CNV. Because VEGFR-2 is overex- and 0.1% trifluoroacetic acid and solution B, 100% acetoni- pressed on neovascular endothelium, normal vessels trile and 0.1% trifluoroacetic acid (gradient starting at solu- should be relatively spared and retinal cells should be un- tions A-B, 80%:20% and going to 80% solution B in 45 minutes). affected after PDT using verteporfin targeted to VEGFR-2. The molecular weight of VEGFR-2–targeted verteporfin was de- Experiments were designed to evaluate the efficacy and termined by mass spectrometry and found to be approxi- selectivity of PDT with VEGFR2-targeted verteporfin in mately 30 kDa. For the remainder of this report, the photo- the rat laser-injury model of CNV. sensitizer dose will be expressed in verteporfin-equivalents (in milligrams divided by the square of the body surface in square meters) as determined by spectrofluorometry using a verteporfin calibration curve. Briefly, a calibration curve correlat- METHODS ing concentration vs spectral emission of liposomal verteporfin was constructed. Spectral emission of a sample of ISOLATION OF FREE VERTEPORFIN VEGFR-2–targeted verteporfin was determined and used to ex- FROM ITS LIPOSOMAL FORMULATION trapolate from the calibration curve its verteporfin content. We recovered verteporfin at a concentration of 2 mg/mL in li- All intermediates (free verteporfin, verteporfin-PVA, and posomal formulation from material prepared for clinical treat- VEGFR-2–targeted verteporfin) were found to have the same ments; leftovers were refrigerated and processed within 2 weeks excitation and emission spectra as the liposomal verteporfin for- to ensure activity.22 Liposomal verteporfin was acidified using mulation as determined by spectrofluorometry and to pre- a 6M hydrochloric acid solution, and separation of organic serve an equivalent in vitro photosensitizing activity (tested in (verteporfin) and aqueous (liposome) layers was achieved us- human umbilical vein endothelial cells) as determined by the tetrazolium salt MMT assay.25 ing dichloromethane (CH2Cl2). After concentrating the solution by means of evaporation, verteporfin was further purified by means of gravity chromatography on silica gel using an elut- EXPERIMENTAL CNV MODEL ing solvent consisting of a 3:1 ratio of CH2Cl2 to methanol. The verteporfin solution was evaporated to dryness and redis- The rat laser-injury model of CNV was modified from earlier re- 26-28 solved in dimethyl sulfoxide. ports and used in our laboratory for PDT. Adult male pigmented rats (Brown-Norway; Charles River Laboratories, Wil- SYNTHESIS OF mington, Mass) were used in the study, and all procedures were VEGFR-2–TARGETED VERTEPORFIN conducted in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Previous studies linking photosensitizers to antibodies sug- Research and the guidelines of the Massachusetts Eye and Ear gested that derivatives of polyvinyl alcohol (PVA; molecular Infirmary (Boston) Animal Care Committee. The rats were anes- weight, 10000-11000 Da; Sigma-Aldrich Corp; St Louis, Mo) thetized for all procedures with an intramuscular injection of 0.2 provide suitable carriers for photosensitizers without jeopar- mL of a 50:50 mixture of ketamine hydrochloride (20 mg/mL) dizing the biological activity of the photosensitizer.23 The pro- and xylazine hydrochloride (100 mg/mL) (both from Phoenix cedure for loading verteporfin on PVA has been described else- Pharmaceutical Inc, St Joseph, Mo). For killing, a mixture con- where.23,24 Briefly, PVA was modified with 2-fluoro-1-methyl sisting of pentobarbital sodium, 390 mg/mL; propylene glycol; pyridinium toluene-4-sulfonate (Sigma-Aldrich Corp) and 1,6- ethanol; and water (Fatal Plus, 650 mg/kg; Vortech Pharmaceu- hexanedimanine (Sigma-Aldrich Corp) to produce side chains ticals, Dearborn, Mich) was given intraperitoneally. containing terminal-free amino groups. Conjugation of modi- The pupils were dilated with 5% phenylephrine hydro- fied PVA with verteporfin was affected by reacting it with a 25- chloride and 0.8% tropicamide, and 4 to 6 photocoagulation fold molar excess of verteporfin in the presence of carbodi- lesions using a Coherent 920 argon dye laser (100-µm spot size; imide as coupling agent in dimethyl sulfoxide. Carrier conjugates 0.1-second duration; 630 nm and 120-160 mW) (Coherent were analyzed by means of high-performance liquid chroma- Medical Laser, Palo Alto, Calif) were delivered between the reti- tography using a column with ultrasphere consisting of 5-µm nal vessels in a peripapillary distribution in each fundus using

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1003

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 a slitlamp delivery system and a cover glass as a contact lens. copy, 1-µm sections were stained with 1% toluidine blue in 1% Production of a bubble at the time of laser confirmed the rup- borate buffer and examined with a Zeiss photomicroscope (Ax- ture of the Bruch membrane. The presence of CNV was con- iophot, Oberkochen, Germany). For electron microscopy, sec- firmed by fluorescein angiography using a TRC-50VT camera tions were stained with a saturated aqueous uranyl acetate so- (Topcon, Paramus, NJ) with images captured on IMAGEnet for lution and Sato lead stain. Sections were viewed with a Windows system (Topcon) after an injection of 1 mL of 1% fluo- transmission electron microscope (Philips CM 10; Royal Phil- rescein sodium (Alcon, Fort Worth, Tex). A choroidal neovas- ips Electronics, Eindhoven, the Netherlands). cular membrane was defined as present if early hyperfluorescence with late leakage was present at the site of the inducing SECOND OUTCOME MEASURE: laser injury as previously described.28 EFFECT ON NORMAL CHOROID AND RETINA

PHOTOSENSITIZERS Grading of sections was performed in a masked fashion (as to dose and photosensitizer) by an experienced grader (N.A.M.) Verteporfin conjugates were tested in eyes with experimental using the following histological grading scheme for PDT ef- CNV and in normal eyes. For targeted PDT, verteporfin-PVA fects on normal choroid and retina modified from Kramer et was bound to the peptide ATWLPPR. The untargeted photo- al6: grade 1 indicates damage in the RPE and photoreceptors, sensitizer conjugate, verteporfin-PVA, was chosen as a con- with occasional pyknosis in the outer nuclear layer (ONL), with trol because both molecules have comparable molecular weights or without choriocapillaris damage; grade 2, choriocapillaris (28 and 30 kDa, respectively). The results with the targeted and closure, RPE and photoreceptor damage, and 10% to 20% pyk- control conjugates were compared with previous results in the nosis in the ONL; grade 3, grade 2 with less than 50% pykno- rat model with PDT using verteporfin (718 Da) in liposomal sis in the ONL; grade 4, grade 3 with greater than 50% pykno- formulation, which is the therapy currently in clinical use.28 sis in the ONL; and grade 5, grade 4 with damage to large choroidal vessels or retinal vessels or inner retinal layers. VEGFR-2–TARGETED VERTEPORFIN AND VERTEPORFIN-PVA ANGIOGRAPHY RESULTS Angiographies with VEGFR-2–targeted verteporfin and verteporfin-PVA were performed using a standard fundus camera but TEMPOROSPATIAL LOCALIZATION with verteporfin-specific filters, with the excitation spectral band OF TARGETED VERTEPORFIN centered at 580 nm and fluorescence detection at 695 nm. Maxi- mal gain settings were required. Photosensitizer doses were given To test the temporospatial localization of targeted and via tail-vein injection, and angiography was performed at a drug control verteporfin conjugates in the retinal and choroi- dose of 12 mg/m2. Conversion to body surface area (in square me- dal circulations and CNV, angiography was performed ters) from weight (in kilograms) was made using a nomogram using targeted verteporfin and verteporfin-PVA at a dose developed by Gilpin.29 Relative fluorescence intensities were de- of 12 mg/m2. At the earliest time points captured (5-10 termined by visual analysis of the angiograms. seconds) after drug injection, fluorescence of targeted and control verteporfin conjugates was already noted in the PDT IN A RAT CNV MODEL choroidal and retinal circulation. Fluorescence intensity peaked in the retinal circulation at approximately 30 Photodynamic therapy was performed on experimental CNV and to 60 seconds and then cleared from retinal circulation areas of normal choroid and retina. Laser light of 689 nm was administered using a diode laser (Coherent Medical Laser, Palo (20-21 minutes), followed by the choroidal circulation Alto, Calif) delivered via a slitlamp adapter (Laserlink; Coherent (35-40 minutes). The drug was retained within the CNV Medical Laser). Laser power at the focal plane was measured with with peak intensity of fluorescence seen at 1 hour a power meter (Coherent Fieldmaster; Coherent, Auburn, Calif). (Figure 1). Verteporfin-PVA exhibited the same tem- The laser spot size was set at 750 µm and was confirmed using a porospatial localization. In contrast, liposomal vertepor- micrometer, and the irradiance used was 600 mW/cm2, which was fin in CNV peaked at 15 to 20 minutes.28 delivered for 17, 42, or 83 seconds to achieve total energy doses 2 of 10, 25, or 50 J/cm , respectively. Activating light fluences were EFFICACY OF PDT based on previous dosimetry established in the rat model of CNV WITH TARGETED VERTEPORFIN for verteporfin PDT.28

FIRST OUTCOME MEASURE: CNV CLOSURE Since angiography demonstrated accumulation of targeted verteporfin and verteporfin-PVA in CNV at 1 hour Fluorescein angiograms were performed at 24 hours after treat- after drug administration, this was the time selected for ment. Closure of CNV was defined by absence of leakage from CNV light application for PDT. Photodynamic therapy was per- compared with the baseline angiogram as previously described. formed using targeted verteporfin or verteporfin-PVA on All angiograms were graded in masked fashion (as to dose and pho- 27 or 32 areas of CNV, respectively. Angiographic clo- tosensitizer) by 2 experienced graders (E.S.G. and J.W.M.). sure of CNV was assessed at 24 hours after PDT and was defined as an absence of leakage from the CNV com- HISTOLOGICAL EVALUATION pared with the pretreatment fluorescein angiogram. The angiograms were graded with masking as to light, drug Eyes were enucleated and the lens and anterior segment were removed. The remaining eyecups were placed in a fixative con- dose, and photosensitizer used. Twenty-four hours after taining 2.5% glutaraldehyde and 2% formaldehyde in 0.1M caco- PDT with verteporfin-PVA, occluded CNV typically showed dylate buffer (pH, 7.4) at 4°C overnight. Tissue samples were a circle of hypofluorescence in the early frames corre- then postfixed in 2% osmium tetroxide, dehydrated in a graded sponding to the treatment spot, with late leakage from in- ethanol series, and embedded in epoxy resin. For light micros- jured RPE and choriocapillaris usually originating at the

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1004

C D

Figure 1. Targeted verteporfin angiography in choroidal neovascularization (CNV). A, A typical fluorescein angiogram performed 2 to 3 weeks after the laser induction of the experimental choroidal neovascular membranes and before photodynamic therapy shows early hyperfluorescence. B, Another view shows late leakage corresponding to CNV. C, Angiography of the same eye using targeted verteporfin (12 mg/m2) shows peak intensity in the retinal vasculature at approximately 30 seconds after injection. D, Peak intensity of targeted verteporfin in CNV at 1 hour after injection (arrowheads).

rim of the treatment spot (Figure 2). Open CNV showed the total number of treated CNVs was small, 100% clo- early hyperfluorescence in the area of CNV and late leak- sure of CNV was achieved using targeted verteporfin with age. The same pattern was observed with PDT with verte- drug doses as low as 3 mg/m2 and a light dose of 10 J/cm2. porfin or with targeted verteporfin irradiated at 15 to 20 Similarly, for verteporfin-PVA, 100% closure was achieved minutes (optimal timing for verteporfin but suboptimal using 3 mg/m2 and a light dose of 25 J/cm2. timing of treatment for targeted verteporfin). In contrast, CNV treated with PDT with targeted verteporfin and ac- HISTOPATHOLOGIC FINDINGS tivating light applied at 1 hour showed no early hypofluo- IN TREATED CNV rescence and only rarely any late hyperfluorescence or leakage. In addition, occluded CNV did not show any All lesions defined as closed, regardless of whether tar- hyperfluorescence or late leakage in the area of CNV, con- geted verteporfin or verteporfin-PVA was used and regard- sistent with occlusion of CNV (Figure 3). less of the light energy dose, shared similar histological fea- Effective CNV closure was demonstrated by fluores- tures. Figure 5 shows a section of a CNV membrane treated cein angiography with targeted verteporfin and verteporfin- with 4.5 mg/m2 of targeted verteporfin and 10 J/cm2 at 24 PVA at all tested photosensitizer doses. The Table and the hours after PDT. This lesion was graded angiographically histogram (Figure 4) summarize the effect of PDT on CNV, as closed. Vessels within the CNV showed vacuolization using different photosensitizer and light doses. Although of endothelial cells and occlusion with platelets, fibrin, and

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1005

C D

Figure 2. Photodynamic therapy (PDT) for choroidal neovascularization (CNV) using verteporfin conjugated to a modified polyvinyl alcohol polymer (PVA). A, Early-phase pretreatment fluorescein angiogram shows hyperfluorescence in areas of CNV. B, Another view shows increasing hyperfluorescence and leakage of CNV. C, Early phase of fluorescein angiogram 24 hours after PDT using verteporfin-PVA with 4.5 mg/m2 and laser fluences of 10 (arrowheads), 25 (arrows), or 50 (asterisk) J/cm2 shows hypofluorescence corresponding to the treatment spots. D, Late-phase fluorescein angiogram shows hyperfluorescence (asterisk) and leakage that originated at the rim of the treatment spots (arrowheads and arrows).

erythrocytes. Extravasated erythrocytes were noted, and verteporfin-PVA at a dose of 2.5 or 4.5 mg/m2 and irra- macrophages were seen within and around the treated CNV diation 1 hour after drug injection using a fluence of 25 complex. Proliferating RPE cells can also be seen surround- or 50 J/cm2 gave similar results. Fluorescein angiogra- ing the CNV complex. Gross disruption of the outer retina, phy 24 hours after PDT showed hypofluorescence in the RPE, and Bruch membrane was generally attributed to the area of treatment in the early phase of the angiogram with laser injury inducing the CNV. late hyperfluorescence (Figure 6A and B). Histological examination revealed occlusion of the choriocapillaris SELECTIVITY: HISTOLOGICAL GRADING with RPE necrosis and pyknosis of the ONL ranging from OF PDT ON NORMAL CHOROID AND RETINA occasional to less than 10%, with mild vacuolization and disarray of the inner and outer segments (Figure 7). The Treatment selectivity was investigated by performing PDT inner retinal layers and larger choroidal vessels, how- in normal retina and choroid, using a qualitative assess- ever, showed no damage, and the lesions were classified ment of angiographic findings after PDT and the histo- as grade 1 damage according to the published scheme. logical grading scheme previously described.6 Photody- Photodynamic therapy for normal retina and cho- namic therapy for normal retina and choroid using roid using targeted verteporfin at a dose of 3 or 4.5 mg/m2

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1006

C D

Figure 3. Photodynamic therapy (PDT) for choroidal neovascularization (CNV) using targeted verteporfin. A, Early-phase pretreatment fluorescein angiogram shows hyperfluorescence in areas of CNV. B, Another view shows increasing hyperfluorescence and leakage from CNV. C, Early-phase fluorescein angiogram 24 hours after PDT using targeted verteporfin with 4.5 mg/m2 and laser fluences of 10 (arrowhead), 25 (arrows), or 50 (asterisk) J/cm2. D, Late-phase fluorescein angiogram shows no hyperfluorescence or leakage from CNV consistent with CNV closure at the treatment spots using 10 (arrowhead) or 25 (arrows) J/cm2. Hyperfluorescence is noted at the edge of a treatment spot treated with 50 J/cm2 (asterisk).

and irradiation 1 hour after drug injection using a fluence of 10, 25, or 50 J/cm2 showed even milder effects Angiographic Grading of CNV Closure 24 Hours After PDT* than the PDT using verteporfin-PVA. Fluorescein angi- Drug Dose† ography performed 24 hours after PDT for normal retina Time of and choroid using targeted verteporfin with light ap- Irradiation 1.5 1.5 3 3 4.5 4.5 plied at 1 hour after photosensitizer injection showed no Light fluence, J/cm2 10 25 10 25 10 25 change in the area of treatment (Figure 6C and D). Le- Targeted verteporfin 1 h 4/6 1/4 4/4 4/4 5/5 4/4 sions were difficult to find by light microscopy, and marker Veteporfin-PVA 1 h 3/6 1/6 2/4 5/5 4/4 7/7 lesions were used to ensure localization. The retina ap- Targeted verteporfin 15 min 5/5 1/5 6/6 5/6 peared normal in almost all respects, with all retinal layers appearing similar to control areas at all doses of tar- Abbreviations: CNV, choroidal neovascularization; PDT, photodynamic geted verteporfin and all light fluences with negligible therapy; targeted verteporfin, veteporfin targeted to CNV by binding to effect on RPE. A few pyknotic nuclei were seen in the vascular endothelial growth factor receptor-2; verteporfin-PVA, ONL at all tested doses, always less than 5%, and usu- verteporfin conjugated to a modified polyvinyl alcohol polymer. *Unless otherwise indicated, data are expressed as number of CNV ally only a few per field (Figure 8). All lesions were clas- lesions closed/number of CNV lesions. sified as grade 1, but showed much less damage than the †Delivered as verteporfin equivalents, calculated as the dose in typical grade 1 lesion because there was minimal effect milligrams divided by the square of the body surface in meters.

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1007

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 on the RPE and virtually none on the photoreceptors. Clo- olization. No changes were seen in the cells and capil- sure of the choriocapillaris was the single consistent laries in the inner retina. marker of PDT damage. Electron microscopy (Figure 9) confirmed the occlusion of the choriocapillaris by plate- COMMENT lets, leukocytes, erythrocytes, and occasionally clumps of fibrin. The choriocapillaris endothelium was usually Photodynamic therapy with verteporfin as currently prac- damaged and often missing entirely. No extravasation of ticed has relative selectivity for CNV based on specific cells or fibrin was seen. This PDT effect contrasted with treatment variables, including drug, light dose, and tim- the condition of the RPE adjacent to the closed chorio- ing of light application.1,2,5,6,30 However, a wealth of ex- capillaris. No necrotic RPE cells were seen; most RPE cells perimental data in tumor animal models indicates that had normal-appearing mitochondria and intact basal in- small molecules, which do not bind specifically to a tu- foldings. Intracellular vacuoles were occasionally seen. mor marker, discriminate poorly between tumor and nor- Rare pyknosis was observed in the ONL, but the inner mal tissues in vivo.31,32 Similarly, preclinical studies have segment mitochondria appeared normal. The outer seg- shown that PDT using untargeted verteporfin for CNV ments showed some disorganization but not much vacu- and normal retina and choroid causes some damage to normal structures. The use of homing vehicles, such as a recombinant monoclonal antibody or a targeting pep- VEGFR-2–Targeted Verteporfin at 1 h tide to VEGFR-2 with the ability to selectively target neo- Verteporfin-PVA at 1 h 120 vascular endothelium, should be useful in improving PDT Verteporfin at 15-20 min outcomes and expanding its applications. Our choice of a peptide as a homing vehicle was mul- 100 tifactorial and included the ease of synthesis and modi- fication, the lack of tissue cross-reactivity, a minimized 80 immunological reaction, a low cost of production, and, most important, the potential incorporation of multiple

60 targeting peptides to the photosensitizer-carrier com-

% Closure plex. This last factor might allow one to target different molecular markers expressed by the pathological tissue 40 to achieve even higher levels of specificity by directing the cytotoxic agent through a distinct routing path to the 20 desired cell or subcellular compartment.33-36 Angiography demonstrated peak localization of the

0 targeted verteporfin and verteporfin-PVA to CNV by 1 1.5 × 10 1.5 × 253 × 10 3 × 25 4.5 × 10 4.5 × 25 6 × 10 6 × 25 hour after intravenous administration, with both drugs 2 × 2 Drug Dose (mg/m ) Fluence (J/cm ) clearing from the CNV by 2 hours. In contrast, lipo- Figure 4. Histogram summarizes the effect of photodynamic therapy on somal verteporfin reaches peak intensity in rat CNV at choroidal neovascularization (CNV) using different photosensitizers and light 15 to 20 minutes after intravenous administration and doses as graded angiographically. All of the drug and light variables tested clears within 30 minutes.28 Targeted verteporfin and verte- for each formulation achieved closure of CNV at some percentage. porfin-PVA are relatively large molecules (28-30 kDa VEGFR-2–targeted verteporfin indicates verteporfin targeted to CNV by binding to vascular endothelial growth factor receptor-2; verteporfin-PVA, compared with 718 Da for unbound verteporfin), and verteporfin conjugated to a modified polyvinyl alcohol polymer. this probably accounts for the later and more prolonged

A B

Figure 5. Light micrographs of choroidal neovascularization 24 hours after photodynamic therapy using verteporfin conjugated to a modified polyvinyl alcohol polymer (A) and targeted verteporfin (B). The treatment variables included a drug dose of 4.5 mg/m2 and a laser fluence of 10 J/cm2. Capillaries (arrowheads) within the lesion are closed (toluidine blue, original magnification ϫ40).

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1008

1:31 6:12

C D

1:19 7:44

Figure 6. A and B, Fluorescein angiography of normal choroid 24 hours after photodynamic therapy (PDT) using verteporfin conjugated to a modified polyvinyl alcohol polymer (4.5 mg/m2) and laser fluence of 25 (arrowhead) or 50 (empty arrowhead) J/cm2 showing early hypofluorescence and mild hyperfluorescence in the late frame (B). Marker laser spots are indicated with arrows. C and D, Fluorescein angiogram of normal choroid 24 hours after PDT treated with targeted verteporfin (4.5 mg/m2) and laser fluence of 25 J/cm2. No hyperfluorescence is noted in the treated area (arrowhead). Marker laser spots (arrows) were used to ensure identification of the treated area.

localization to CNV. To lentino and colleagues37 dem- drugs differed somewhat in their selectivity. Angiography onstrated localization of fluoresceinated dextrans and an- after PDT of normal retina and choroid using verteporfin- tibodies in experimental CNV after intravenous admin- PVA demonstrated early hypofluorescence and late leak- istration and a correlation with molecular weight and age 24 hours after PDT, and results of the histological ex- radius. Presumably, molecules of this size are able to exit amination showed grade 1 lesions with RPE necrosis, mild the vascular space through fenestrations in the chorio- pyknosis of the photoreceptor nuclei, and vacuolization and capillaris and CNV, but are less readily cleared than are disarray of the inner and outer segments. Although this dam- smaller molecules. Thus it is not surprising that PDT us- age is still mild, it is no better than PDT with standard verte- ing both larger molecules, targeted verteporfin and verteporfin. One can speculate that the verteporfin-PVA can still porfin-PVA, showed greater efficacy for CNV closure with leak through the CNV without binding to the endothe- lower drug and light doses than those seen with PDT us- lium, reaching the extravascular space and clearing slowly. ing unbound verteporfin. One hundred percent closure In contrast, PDT for normal retina and choroid using tar- was achieved with as little as 3 mg/m2 and 10 J/cm2 for geted verteporfin did not show any angiographic hypo- targeted verteporfin and 25 J/cm2 for verteporfin-PVA, fluorescence or late leakage 24 hours after PDT, and re- although even with 3 or 6 mg/m2 and 25 J/cm2, the clo- sults of the histological examination showed minimal effect sure rate ranges from 83% to 92%.28 on the RPE and no injury to photoreceptors. Although these Although the efficacy of CNV closure was similar for lesions were formally classified as grade 1 lesions, the ob- PDT using targeted verteporfin and verteporfin-PVA, the served damage was substantially less. Presumably, the

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1009

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 VEGFR-2–targeted verteporfin localizes to CNV on the ba- CONCLUSIONS sis of size in a manner similar to verteporfin-PVA, but can then bind VEGFR-2 receptors expressed on neovascular We showed that the efficacy of PDT can be enhanced by endothelium. This binding could lead to increased effi- conjugating the photosensitizer to PVA. In addition, se- cacy of CNV closure and increased selectivity, because the lectivity of PDT can be enhanced by targeting the pho- photosensitizer would be sequestered at the endothelium tosensitizer to VEGFR-2. If these preliminary findings are and would spare the RPE and photoreceptors. substantiated in primate models of CNV, clinical stud- An additional advantage to PDT using VEGFR-2–tar- ies may be warranted to determine whether vision out- geted verteporfin might be the combination of an antian- comes can be improved. Our results also highlight the 9,10 giogenic effect with PDT. Previous studies from our group utility of designing peptide photosensitizer conjugates as have shown that combining antiangiogenic agents with PDT vehicles for regulating the distribution of photosensi- causes a selective increased cytotoxicity to neovascular endothelial cells in vitro and in vivo. The VEGFR-2–targeting peptide used in the present study has been shown by Bin- etruy-Tournaire and colleagues21 to completely inhibit VEGF-induced angiogenesis. Thus, VEGFR-2–targeted verteporfin has the potential to exert an antiangiogenic po- tentiating effect before its photoactivation. m

ONL p

OS r

Figure 7. Light microscopy of normal choroid 24 hours after photodynamic therapy using verteporfin conjugated to a modified polyvinyl alcohol polymer (4.5 mg/m2) and laser fluence of 50 J/cm2. The choriocapillaris (cc) was occluded by thrombus, and the retinal pigment epithelium was necrotic. Figure 9. Electron microscopy of normal choroid 24 hours after Vacuolization and disarray of the inner and outer segments (OS) were seen. photodynamic therapy using targeted verteporfin (4.5 mg/m2) and laser Pyknosis of the photoreceptor nuclei (outer nuclear layer [ONL]) was less fluence of 50 J/cm2. There is closure of choriocapillaris with red blood cells than 10%, and the inner retina appeared intact. This lesion was classified as (r), platelets (p), and fibrin in the lumen. The endothelium is missing (arrow). having grade 1 damage. Arrow marks the level of the Bruch membrane Mitochondria (m) in the retinal pigment epithelium appear normal, and (toluidine blue, original magnification ϫ40). well-preserved basal infoldings remain (asterisk). Bar indicates 1 µm.

A B

OS ONL

RPE OS CC

Figure 8. Light microscopy of normal choroid 24 hours after photodynamic therapy using targeted verteporfin (4.5 mg/m2) and laser fluence of 50 J/cm2.A,The retinal layers appear to be well preserved. Occasional pyknotic nuclei are seen in the outer nuclear layer (ONL) as indicated by arrowheads (original magnification ϫ40). The Bruch membrane is marked by the arrow. The lesion was classified as grade 1 damage. B, There is closure of choriocapillaris (cc) with well-preserved overlying layer of retinal pigment epithelium (RPE). OS indicates outer segments (original magnification ϫ100).

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1010

©2004 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 tizer to CNV to maximize their selectivity in PDT. In the retreatment of normal retina and choroid in the cynomolgus monkey. Ophthal- mology. 1999;106:1915-1923. future, other candidate homing molecules may be iden- 9. Renno R, Delori F, Holzer R, Gragoudas E, Miller J. Photodynamic therapy using tified with even greater specificity for neovascular endo- Lu-Tex induces apoptosis in vitro and shows potentiate action combined with thelium. These preliminary results indicate that tar- angiostatin in retinal capillary endothelial cells. Invest Ophthalmol Vis Sci. 2000; 41:3963-3971. geted PDT for CNV is feasible and may offer a qualitative 10. Gauthier D, Husain D, Kim I, et al. Safety and Efficacy of Intravitreal Injection of improvement in current clinical therapies. rhuFab VEGF in Combination With Verteporfin PDT on Experimental Choroidal Neovascularization. Fort Lauderdale, Fla: Association for Research in Vision & Ophthalmology; 2002. Submitted for publication August 12, 2002; final revision 11. Pasqualini R, Ruoslahti E. Organ targeting in vivo using phage display peptide received October 23, 2003; accepted January 30, 2004. libraries. Nature. 1996;380:364-366. 12. Arap W, Pasqualini R, Ruoslahti E. Cancer treatment by targeted drug delivery From the Angiogenesis and Laser Laboratory, Retina to tumor vasculature in a mouse model. Science. 1998;279:377-380. Service, Massachusetts Eye and Ear Infirmary, Depart- 13. Arap WHW, Bernasconi M, Kain R, et al. Targeting the prostate for destruction ment of Ophthalmology, Harvard Medical School, Boston through a vascular address. Proc Natl Acad Sci U S A. 2002;99:1527-1531. 14. Miller J, Adamis A, Shima D, et al. Vascular endothelial growth factor/vascular (Drs Renno, Terada, Gragoudas, and Miller and Mr Mi- permeability factor is temporally and spatially correlated with ocular angiogen- chaud), and the Department of Chemistry, American Uni- esis in a primate model. Am J Pathol. 1994;145:574-584. 15. Aiello L, Pierce E, Foley E, et al. Suppression of retinal neovascularization in vivo versity of Beirut, Beirut, Lebanon (Dr Haddadin). Dr Renno by inhibition of vascular endothelial growth factor (VEGF) using soluble VEGF- is now with the Department of Ophthalmology, Jules Stein receptor chimeric proteins. Proc Natl Acad Sci U S A. 1995;92:10457-10461. Eye Institute, University of California–Los Angeles. The Mas- 16. McLeod D, Taomoto M, Cao J, Zhu Z, Witte L, Lutty G. Localization of VEGF receptor-2 (KDR/Flk-1) and effects of blocking it in oxygen-induced retinopathy. sachusetts Eye and Ear Infirmary has an ownership inter- Invest Ophthalmol Vis Sci. 2002;43:474-482. est in 3 US patents directed to the use of verteporfin. In ad- 17. Tille J, Wood J, Mandriota S, et al. Vascular endothelial growth factor (VEGF) dition, the Massachusetts Eye and Ear Infirmary has an receptor-2 antagonists inhibit VEGF- and basic fibroblast growth factor– induced angiogenesis in vivo and in vitro. J Pharmacol Exp Ther. 2001;299: ownership interest in certain patent applications directed 1073-1085. to the selective destruction of subretinal choroidal neovas- 18. Krzystolik M, Afshari M, Adamis A, et al. Prevention of experimental choroidal neovascularization with intravitreal anti–vascular endothelial growth factor an- culature for the treatment of macular degeneration and other tibody fragment. Arch Ophthalmol. 2002;120:338-346. disorders. Should the Massachusetts Eye and Ear Infir- 19. Kim I, Ryan A, Rohan R, et al. Constitutive expression of VEGF, VEGFR-1, and mary receive royalties or other financial remuneration as VEGFR-2 in normal eyes. Invest Ophthalmol Vis Sci. 1999;40:2115-2121. 20. Wada M, Ogata N, Otsuji T, Uyama M. Expression of vascular endothelial growth a result of these patents and patent applications, Drs Miller, factor and its receptor (KDR/flk-1) mRNA in experimental choroidal neovascu- Renno, and Gragoudas would receive a share of the same larization. Curr Eye Res. 1999;18:203-213. in accordance with the Massachusetts Eye and Ear Infir- 21. Binetruy-Tournaire R, Demangel C, Malavaud B, et al. Identification of a peptide blocking vascular endothelial growth factor (VEGF)–mediated angiogenesis. EMBO mary’s institutional patent policy and procedures, which in- J. 2000;19:1525-1533. cludes royalty-sharing provisions. 22. Lange N, Ballini J, Wagnieres G, Bergh HVD. A new drug-screening procedure for photosensitizing agents used in photodynamic therapy for CNV. Invest Oph- This study was supported by the Foundation Fighting thalmol Vis Sci. 2001;42:38-46. Blindness, Owings Mills, Md (Drs Renno, Terada, and 23. Steele K, Liu D, Davis N, Deal H, Levy J. The preparation and application of porphyrin- Miller), and the Iacocca Foundation, Boston, Mass (Dr monocalonal antibodies for cancer therapy. In: Dougherty TJ, ed. Photodynamic Therapy: Mechanisms: 19-20 January 1989, Los Angeles, California. Vol 1065. Bel- Renno). lingham, Wash: International Society for Optical Engineering; 1989:73-79. Correspondence: Joan W. Miller, MD, Angiogenesis and 24. Jiang FN, Jiang S, Liu D, Richter A, Levy JG. Development of technology for link- Laser Research Laboratory, Retina Service, Massachusetts ing photosensitizers to a model monoclonal antibody. J Immunol Methods. 1990; 134:139-149. Eye and Ear Infirmary, 243 Charles St, Boston, MA 02114 25. Monner D. An assay for growth of mouse bone marrow cells in microtiter liquid ([email protected]). culture using the tetrazolium salt MTT, and its application to studies of myelo- poresis. Immunol Lett. 1988;19:261-268. 26. Dobi E, Puliafito C, Destro M. A new model of subretinal neovascularization in REFERENCES the pigmented rat. Arch Ophthalmol. 1989;107:264-269. 27. Tobe T, Takahashi K, Ohkuma H, Uyamam M. Experimental choroidal neovascularization in the rat [in Japanese]. Nippon Ganka Gakkai Zasshi. 1994;98:837-845. 1. Miller J, Schmidt-Erfurth U, Sickenberg M, et al. Photodynamic therapy for cho- 28. Zacks D, Ezra E, Terada Y, et al. Verteporfin photodynamic therapy in the rat model roidal neovascularization due to age-related macular degeneration with verte- of choroidal neovascularization: angiographic and histologic characterization. In- porfin: results of a single treatment in a phase 1 and 2 study [published correc- vest Ophthalmol Vis Sci. 2002;43:2384-2391. tion appears in Arch Ophthalmol. 2000;118:488]. Arch Ophthalmol. 1999;117: 29. Gilpin D. Calculation of a new Meeh constant and experimental determination of 1161-1173. burn size. Burns. 1996;22:607-611. 2. Treatment of Age-Related Macular Degeneration With Photodynamic Therapy (TAP) 30. Husain D, Miller J, Michaud N, Connolly E, Flotte T, Gragoudas E. Intravenous Study Group. Photodynamic therapy of subfoveal choroidal neovascularization infusion of liposomal benzoporphyrin derivative for photodynamic therapy of ex- in age-related macular degeneration with verteporfin: one-year results of 2 ran- perimental choroidal neovascularization. Arch Ophthalmol. 1996;114:978-985. domized clinical trials: TAP report 1 [published correction appears in Arch Oph- 31. Murdter TE, Sperker B, Kivisto KT, et al. Enhanced uptake of doxorubicin into bron- thalmol. 2000;118:488]. Arch Ophthalmol. 1999;117:1329-1345. chial carcinoma: beta-glucuronidase mediates release of doxorubicin from a gluc- 3. Bressler N; Treatment of Age-Related Macular Degeneration With Photodynamic uronide prodrug (HMR 1826) at the tumor site. Cancer Res. 1997;57:2440-2445. Therapy (TAP) Study Group. Photodynamic therapy of subfoveal choroidal neovas- 32. Folli S, Westermann P, Braichotte D, et al. Antibody-indocyanin conjugates for cularization in related age macular degeneration with verteporfin: two-year results immunophotodetection of human squamous cell carcinoma in nude mice. Can- of randomized clinical trials: TAP report 2. Arch Ophthalmol. 2001;119:198-207. cer Res. 1994;54:2643-2649. 4. Sickenberg M, Schmidt-Erfurth U, Miller J, et al. A preliminary study of photo- 33. Vocero-Akbani AM, Heyden NV, Lissy NA, Ratner L, Dowdy SF. Killing HIV- dynamic therapy using verteporfin for choroidal neovascularization in patho- infected cells by transduction with an HIV protease-activated caspase-3 protein. logic myopia, ocular histoplasmosis syndrome, angioid streaks, and idiopathic Nat Med. 1999;5:29-33. causes. Arch Ophthalmol. 2000;118:327-336. 34. Schwarze SR, Ho A, Vocero-Akbani A, Dowdy SF. In vivo protein transduction: 5. Miller J, Walsh A, Kramer M, et al. Photodynamic therapy of experimental cho- delivery of a biologically active protein into the mouse. Science. 1999;285:1569- roidal neovascularization using lipoprotein-delivered benzoporphyrin. Arch Oph- 1572. thalmol. 1995;113:810-818. 35. Schwarze SR, Dowdy SF. In vivo protein transduction: intracellular delivery of 6. Kramer M, Miller J, Mihaud N, et al. Liposomal benzoporphyrin derivative verte- biologically active proteins, compounds and DNA. Trends Pharmacol Sci. 2000; porfin photodynamic therapy. Ophthalmology. 1996;103:427-438. 21:45-48. 7. Husain D, Kramer M, Kenney A, et al. Effects of photodynamic therapy using verte- 36. Lindgren M, Hallbrink M, Prochiantz A, Langel U. Cell-penetrating peptides. Trends porfin on experimental choroidal neovascularization and normal retina and cho- Pharmacol Sci. 2000;21:99-103. roid up to seven weeks after treatment. Invest Ophthalmol Vis Sci. 1999;40: 37. Tolentino M, Husain D, Theodosiadis P, et al. Angiography of fluoresceinated 2322-2331. anti-vascular endothelial growth factor antibody and dextrans in experimental 8. Reinke M, Canakis C, Hussain D, et al. Verteporfin photodynamic therapy (PDT) choroidal neovascularization. Arch Ophthalmol. 2000;118:78-84.

(REPRINTED) ARCH OPHTHALMOL / VOL 122, JULY 2004 WWW.ARCHOPHTHALMOL.COM 1011