Ocular Inflammation in Cynomolgus Macaques Following Intravenous Administration of a Human Monoclonal Antibody

Abstract

Angiogenesis is a major component of the pathogenesis of various ocular diseases, including age-related macular degeneration (AMD). CNTO95 is a fully human monoclonal antibody against α_ν integrins that has shown antiangiogenic properties in cynomolgus macaques and rats. Because angiogenesis inhibitors may have the potential to treat AMD, a proof-of-concept study was conducted in a macaque model of laser-induced choroidal neovascularization. In the course of this study, transient, intense anterior chamber ocular inflammation was observed within 24 hours following the first intravitreal or intravenous administration of the human monoclonal antibody. These animals had no outward signs of ocular toxicity or discomfort. Additional ocular safety studies demonstrated that the inflammation following intravenous administration of CNTO95 was not due to a contaminant in the vehicle, not due to endotoxin, and not a nonspecific reaction in the macaques from administration of a human monoclonal antibody. The anterior chamber ocular inflammation noted following the first dose did not recur with subsequent CNTO95 dosing. In repeated-dose toxicology studies, histopathological examination of the eyes revealed no ocular toxicity. The reason for the ocular inflammation following intravenous dosing remains unresolved but may be a secondary manifestation of a first-dose systemic infusion reaction.

Angiogenesis, the formation and differentiation of new blood vessels, is essential to physiologic processes such as growth, reproduction, and wound healing. ¹ Angiogenesis can, however, become deregulated in certain pathologic processes, such as cancer and various ocular diseases, including age-related macular degeneration (AMD) and proliferative retinopathy. ^2–4

Vascular endothelial growth factor (VEGF) induces angiogenesis and increases vascular permeability and is an important mediator of tumor angiogenesis and ocular neovascular diseases. ^4–6 The anti-VEGF monoclonal antibody, bevacizumab (Avastin), is currently approved for the treatment of various tumors. The anti-VEGF antibody fragment, ranibizumab (Lucentis), and the anti-VEGF oligonucleotide, pegaptanib (Macugen), are currently approved for neovascular AMD.

Integrins have also been shown to be essential components of angiogenesis. ^7,8 Integrins facilitate the adhesion of stimulated endothelial cells to the extracellular matrix, trigger the secretion of extracellular matrix-rearranging proteases, and propagate signaling events that promote the survival and differentiation of the cells in the newly formed vasculature.

CNTO95 is a fully human monoclonal antibody that binds to α_νβ₃, α_νβ₅, α_νβ₁, and α_νβ₆ integrins; inhibits adhesion of endothelial cells to the extracellular matrix in vitro; inhibits angiogenesis in vivo; and inhibits human tumor growth in rodents. ^9,10 CNTO95 is currently in clinical trials for the treatment of cancer. A phase I clinical study has been completed with CNTO95 in patients with advanced solid tumors. ¹¹ In that study, the most frequently observed adverse events were fever, rigors, and headache, which occurred within the first 8 hours of the infusions. Infusion-related reactions have been reported for a number of monoclonal antibody therapeutics. ^12,13 In many cases, the infusion reactions are more intense following the first infusion than following subsequent infusions and are therefore referred to as “first-dose” infusion reactions.

To support the dosing of CNTO95 to patients in clinical trials, repeated-dose toxicology studies have been conducted in cynomolgus macaques. ¹⁴ The cynomolgus macaque is a pharmacologically relevant species for the evaluation of CNTO95 safety and efficacy. CNTO95 has been shown to bind to macaque aortic endothelial cells with an affinity similar to that of human cells and to inhibit angiogenesis in subcutaneously implanted basic fibroblast growth factor (bFGF) containing matrigel plugs in cynomolgus macaques. ⁹ The toxicology studies conducted in cynomolgus macaques to support clinical dosing with CNTO95 showed no adverse effects. However, in the course of a pharmacology study designed to evaluate the potential of CNTO95 to inhibit laser-induced choroidal neovacularization (CNV) in macaques (a model for AMD), an unexpected observation of ocular inflammation was observed following intravenous administration of this antibody. These studies describe the ocular inflammation that was observed and the additional studies designed to further elucidate the nature of the inflammation. The reason for the inflammation remains unsolved but may be a secondary manifestation of a more generalized first-dose infusion reaction.

Materials and Methods

CNTO95 is a fully human IgG1 monoclonal antibody (mAb) developed and manufactured by Centocor, R&D, Inc (Horsham, Pennsylvania). The CNTO95 used in these studies was formulated at a concentration of 20 mg/mL in a buffer containing 0.01 M sodium phosphate, 10% sucrose, and 0.001% to 0.012% polysorbate 80 (pH 5.5–6.5). The endotoxin concentration in the CNTO95 product was less than 0.03 EU/mg. The human anti-HER2/neu mAb, trastuzumab (Herceptin), and the human anti-VEGF mAb, bevacizumab (Avastin), were purchased from commercial pharmacies. The endotoxin concentrations in the commercial, clinical-grade products were less than 0.1 EU/mg.

These studies were conducted in young healthy cynomolgus macaques (Macaca fascicularis). All in-life procedures were conducted in compliance with the Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals, and the Office of Laboratory Animal Welfare. Study protocols were reviewed and approved by the Institutional Animal Care and Use Committees. All histopathological evaluations were conducted by American College of Veterinary Pathologists (ACVP) board-certified veterinary pathologists. All ophthalmic examinations (except for study 7) were conducted by American College of Veterinary Ophthalmologists (ACVO) board-certified veterinary ophthalmologists. The ophthalmic examinations included a slit-lamp biomicroscope examination of the adnexa and anterior portion of the eyes and an indirect ophthalmoscope examination of the ocular fundus of the eyes.

General Toxicity Studies Conducted in Cynomolgus Macaques With CNTO95

To support clinical dosing of CNTO95 in phase I and phase II clinical trials in oncology patients, we conducted 3 repeated-dose toxicology studies (studies 1–3, Table 1). In these studies, saline or CNTO95 (10 or 50 mg/kg) was administered intravenously once per week to cynomolgus macaques (3–6 per sex per treatment group) for 4 to 26 weeks. These studies have been described in detail elsewhere ¹⁴ and are referred to here only in the context of the ocular evaluations that were conducted as part of the general toxicity assessments. Ophthalmic examinations were conducted prestudy and after animals had received at least 4 weekly doses of CNTO95 (during weeks 4 and 8 for study 1, during week 8 for study 2, and during weeks 12, 25, and 36 for study 3). At the end of the study periods, a comprehensive anatomic pathology examination was conducted. A subset of animals in studies 1 and 3 was allowed a treatment-free period prior to necropsy. At necropsy, eyes were collected from all animals and preserved in Davidson’s fixative for optimum fixation. The eyes were embedded in paraffin, sectioned, stained with hematoxylin and eosin, and examined by light microscopy.

Laser-Induced CNV Model

The model of laser-induced CNV used in this study has been described previously ¹⁵ and is similar to the model described by Krzystolik et al. ¹⁶ In both models, intravitreal administration of an anti-VEGF antibody Fab fragment was shown to reduce CNV.

For the evaluation of the effects of CNTO95 on laser-induced CNV, CNTO95 was administered either intravenously at a dose of 50 mg/kg or intravitreally at a dose of 500 μg per eye (50 μL; study 4, Table 1). The doses of 50 mg/kg intravenously and 500 μg per eye were selected to produce serum and vitreal concentrations of CNTO95, respectively, that are sufficient to fully saturate α_ν-integrin receptor binding. ¹⁴ Five male cynomolgus macaques weighing 3.2 to 5.8 kg received slow bolus intravenous injections of CNTO95. The animals in the intravenous treatment group were not injected intravitreally. Eight male cynomolgus macaques weighing 3.6 to 5.1 kg received intravitreal injections of CNTO95 into the right eye and an intravitreal injection of an equivalent volume of saline (50 μL) in the left eye. In these animals, the left eye was used as the control. An additional 8 cynomolgus macaques weighing 3.4 to 4.2 kg received a saline injection into the left eye and were used throughout the study as controls only. Animals were administered CNTO95 by intravitreal or intravenous injection or saline by intravitreal injection on days 1, 15, 29, and 43.

For the intravitreal injections, animals were anesthetized with ketamine and xylazine, and the eyes were rinsed with sterile balanced salt solution. A topical anesthetic (0.5% proparacaine) was instilled in each eye before dose administration. The eyes were cleaned with a 25% betadine solution prepared with 0.9% sodium chloride for injection, USP (2.5% providone iodine) and then rinsed with a physiological salt solution. Eyes were injected intravitreally using a 1-cc tuberculin syringe and 30-gauge needle. A topical antibiotic (Tobrex) was instilled in each eye after dose administration. Animals were observed cage side throughout the study period for any clinical signs, with particular attention to the eyes.

On day 8 or 9, the macula of each eye underwent laser treatment using methods similar to those described previously. ^15–17 Color ocular photographs were taken on the day of laser induction (after the laser treatment) and on day 41. The color ocular photographs included the retina and pertinent ocular abnormalities, stereoscopic photographs of the posterior pole, and nonstereoscopic photographs of 2 mid-peripheral fields (temporal and nasal).

Fluorescein angiography was conducted on days 22, 27, and 41. Animals were anesthetized with ketamine and maintained on isoflurane, and the eyes were dilated with a mydriatic agent. Animals were given an intravenous injection of fluorescein.

Ophthalmic examinations were conducted before the initiation of treatment, 1 day after the first treatment (day 2), 5 days after the first treatment (day 6), and 1 day after the last (fourth biweekly) treatment (day 44). Animals were anesthetized with ketamine, and the eyes were dilated with a mydriatic agent. The adnexa and anterior portion of both eyes was examined using a handheld slit-lamp biomicroscope. The ocular fundus of both eyes was examined using an indirect ophthalmoscope. Aqueous and vitreal cell scores were assigned using the same estimate of cells per single 0.2-mm field of the focused slit-lamp beam (0 = no cells, trace = 1-5 cells, 1+ = 5–25 cells, 2+ = 25–50 cells, 3+ = 50–100 cells, and 4+ = more than 100 cells). Aqueous flare was scored as follows: 0 = no visible protein in anterior chamber; trace = anterior chamber protein is visible only to an experienced observer using a small, bright focal light source and magnification; 1+ = mild; 2+ to 3+ = moderate protein, with 3+ more severe than 2+ ; and 4+ = severe (modified from Hogan et al ¹⁸ ).

On day 44, serum and vitreal samples were collected for analysis of CNTO95 concentration and were analyzed using methods described previously. ¹⁴

Ocular Inflammation Follow-Up Studies

A follow-up study was conducted to identify a noeffect dose level for clinical ophthalmic findings following intravenous administration of CNTO95 (study 5, Table 1). In this study, the effects of intravenously administered CNTO95 were compared with a saline control, vehicle control, and an isotype (human IgG1)–matched mAb (trastuzumab) control.

Male and female cynomolgus macaques weighing 2.1 to 3.9 kg were assigned to 5 groups of 3 animals per group. Animals were dosed by bolus intravenous injection on days 1, 8, 15, and 22. Group 1 animals received saline control (0.9% sodium chloride for injection, USP; sterile saline) on days 1 and 8 and vehicle control (0.01 M sodium phosphate, 10% (w/v) sucrose [pH 6], 0.0001% polysorbate 80) on day 15. On day 22, group 1 animals received CNTO95 (50 mg/kg). Groups 2, 3, and 4 received CNTO95 at 2, 10, or 50 mg/kg, respectively (except group 4 animals on day 22 received CNTO95 vehicle). Group 5 received trastuzumab at 50 mg/kg. Doses were administered at volumes ranging from 0.1 to 2.5 mL/kg.

Ophthalmic examinations were conducted prior to initiation of dosing, on the day following dosing, and 7 days after each dose. Body temperature and hematology evaluations were conducted on the same days as the ophthalmic examinations. Clinical observations and body weight were evaluated throughout the study. Animals were returned to the colony at the end of the observation periods.

An additional 9 male cynomolgus macaques weighing 2.5 to 2.8 kg received a single intravenous bolus injection or 2-hour infusion of CNTO95 (50 mg/kg) on day 1 (study 6, Table 1). Ophthalmic examinations were conducted on days 2, 3, 4, 5, and 8 to evaluate the time course and resolution of the ocular inflammation. Animals were returned to the colony at the end of the observation periods.

Additional Toxicology Studies

In 2 additional studies, ophthalmic examinations were conducted in cynomolgus macaques at 24 hours after the first administration of CNTO95 (studies 7 and 8, Table 1). In study 7, 2 male (3.9–4 kg) and 1 female (2.5 kg) cynomolgus macaques received a slow bolus intravenous injection of CNTO95 (10 mg/kg) on day 1 followed by ophthalmic examinations (by the staff veterinarian) on day 2. In study 8, CNTO95 (20 mg/kg) or saline was administered intravenously to 3 male (2.6–3 kg) and 3 female (3.6–4.5 kg) cynomolgus macaques per treatment group once per week for 4 weeks. This study also included a group of 3 male (2.8–3 kg) and 3 female (2.7–3.5) animals that received bevacizumab (anti-VEGF mAb) 20 mg/kg once per week for 4 weeks. The purpose of the bevacizumab group was to compare the toxicity of CNTO95 to that of a second antiangiogenesis inhibitor. Ophthalmic examinations were conducted 24 hours after the first dose, and histopathological examinations were conducted at the end of the 4-week treatment period.

Immunolocalization of CNTO95 in Macaque and Human Eye Tissues In Vitro

The in vitro binding of CNTO95 to cryosections of macaque and human eye tissues was evaluated. Human placenta was included as a positive control tissue. Eye tissues were obtained at necropsy from untreated cynomolgus macaques or at autopsy from normal human donors. Fresh unfixed tissue samples were placed in molds and frozen on dry ice in OCT embedding medium. Tissues were maintained below −70°C until sectioning. Sections were cut at approximately 5 μm and fixed in room temperature in acetone for 10 minutes. Just prior to staining, the slides were fixed in 10% neutral buffered formalin for 10 seconds. An indirect immunoperoxidase procedure was performed to evaluate CNTO95 localization using methods described previously. ¹⁴ Briefly, cryosections were blocked sequentially with avidin, biotin, and a protein block designed to reduce nonspecific binding. CNTO95 was applied to the tissues in vitro at a concentration of 10 μg/mL. CNTO95 binding was detected using a mouse antiidiotypic antibody toward CNTO95 followed by goat antimouse antibody and then avidin-biotin complex. All slides were counterstained with hematoxylin, dehydrated, and coverslipped for interpretation. Slides were examined by light microscopy and grade using the following scale: 0 = negative, ± = equivocal, 1+ = weak, 2+ = moderate, 3+ = strong, and 4+ = intense.

Serum and Vitreous CNTO95 Concentration Analysis

Serum and vitreous CNTO95 concentrations were measured using a validated sandwich enzyme immunoassay (EIA). The limit of quantification of the assay was 0.05 μg/mL (for a 1/10 dilution) of CNTO95 in the matrix. Two different murine antiidiotypic antibodies against CNTO95 were used for this assay, one for capture and one for detection. Animal serum samples diluted in phosphate-buffered saline (PBS), and 1% bovine serum albumin (BSA) was added to the capture antibody-coated plates in triplicate. Biotin-coupled detection antibody diluted in BSA + PBS buffer was used as the secondary antibody followed by streptavidin-conjugated horseradish peroxidase (Jackson Immunoresearch Laboratories, West Grove, Pennsylvania). Plates were washed and tetramethylbenzidine (Kirkegaard and Perry Labs, Gaithersburg, Maryland) substrate was added. The colorimetric reaction was stopped with 4N H₂SO₄. The 450- to 650-nm OD values of the wells were determined using a spectrophotometric microplate reader. The concentration of each unknown was calculated from a standard curve.

Results

General Toxicity Studies

The histopathological examination of eyes from macaques treated with CNTO95 or bevacizumab for 4 to 26 weeks (studies 1, 2, 3, and 8) did not reveal any treatment-related ocular toxicities. In these repeated-dose toxicology studies, the ophthalmic examinations that were conducted following the first-dose administration in 1 study (study 8) and following multiple-dose administration in the other 3 studies (studies 1–3) revealed no ophthalmologic findings (refer to Table 1).

Laser-Induced CNV Model

Administration of CNTO95 either intravenously or intravitreally at the described dose levels did not inhibit laser-induced CNV in this model (study 4; results not shown). No evidence of retinal toxicity was apparent by clinical examination or by evaluation of either the color photographs or the fluorescein angiograms.

Ophthalmic examinations indicated that CNTO95, administered intravitreally at a dose of 500 μg or intravenously at a dose of 50 mg/kg, produced an acute (24 hours postdose), intense inflammatory response in the anterior chamber of the eyes (Figure 1). Inflammatory cells in the anterior chamber were graded as 3 to 4+, and aqueous flare was graded as 1 to 4+. The anterior segment inflammation was largely resolved by day 6, at which time inflammatory cells could be visualized in the vitreous. The inflammation observed in CNTO95-treated animals on days 2 and 6 occurred prior to the laser-induced injury, which was induced on day 8 or 9. Additional administration of CNTO95 on days 15, 29, and 43 resulted in a greatly diminished anterior chamber inflammatory response, even after laser-induced injury (day 8 or 9), so that by 24 hours after the fourth injection (day 44, the next ophthalmic examination time point), the inflammatory response was comparable to that of saline control eyes. In the animals that received intravitreal injections of saline on days 1, 15, 29, and 43, 0 to trace inflammatory cells were detected in the anterior chamber on days 2, 6, and 44.

Posterior segment inflammation in the vitreous exhibited a similar pattern to that seen in the anterior segment, although, as expected, inflammatory cells in the vitreous were slower to appear.

Conjuctival hyperemia was seen only in eyes injected intravitreally with CNTO95 or saline. One day after the first injection, the hyperemia was graded as moderate (2+) in the CNTO95-injected eyes and trace in the saline-treated eyes. One day after the fourth injection, hyperemia was graded as mild in all injected eyes.

Two of 8 eyes receiving CNTO95 intravitreally and 2 of 10 eyes receiving CNTO95 intravenously had corneal keratic precipitates 1 day after the first dose. Keratic precipitates were not present on ophthalmic examinations conducted on days 6 and 44.

In the animals receiving CNTO95 intravenously, there were no outward signs of toxicity; pupillary miosis was not observed, the conjunctiva was not hyperemic, and there were no other outward manifestations of ocular inflammation (eg, blepharospasm, tearing, or other signs of discomfort).

Follow-Up Studies Evaluating Ocular Inflammation Following Intravenous Injection of CNTO95

Follow-up studies were conducted to further evaluate the nature of the ocular inflammation observed following intravenous dosing of CNTO95 (studies 5 and 6).

In study 5, intravenous administration of CNTO95 or an isotype-matched human monoclonal antibody (anti-HER2 mAb, traztuzumab) had no effect on clinical observations, body weight, body temperature, or hematology evaluations. Administration of CNTO95 resulted in intraocular inflammation in all treated animals on the day following dosing (∼24 hours postdose; Figure 2). Overall, this reaction was less in the 2-mg/kg treatment group than in the 10-mg/kg and 50-mg/kg treatment groups, but the small sample size precluded making unequivocal dose–response comparisons, and 1 or more animals in each of the 3 groups exhibited a considerable inflammatory response (2 to 3+ aqueous flare/4+ anterior chamber cell). This inflammatory response rapidly and spontaneously improved. By day 8, all eyes in all 3 CNTO95–treated groups had either no or only trace anterior chamber cell responses. With additional doses of CNTO95, the anterior chamber inflammatory response was markedly diminished, and all 3 CNTO95-treated groups exhibited almost no or only trace anterior chamber cell responses on all examinations after day 2. Vitreal cell scores tended to parallel those in the anterior chamber, although the peak vitreal cell scores were typically not achieved until 1 week after inflammatory cells first appeared in the anterior chamber. Vitreal cells also tended to be slower to resolve than those in the aqueous humor. Many of the vitreal cells after day 8 were brown in color, suggesting that they represented previous inflammation rather than persistent active inflammation. Group 1 animals that received CNTO95 on day 22 only showed ophthalmic observations on day 23 similar to those seen in group 4 animals on day 2 (1 day after the first dose).

Additional findings on day 2 included the deposition of a very small clump of pigment on the corneal endothelium of 1 eye of 1 animal in the 2-mg/kg group and the presence of white, inflammatory keratic precipitates on the corneal endothelium of both eyes of 2 animals in the 10-mg/kg group and 1 animal in the 50-mg/kg group. These deposits were secondary to anterior segment inflammation, and only the minute pigment clump persisted past day 2. In all CNTO95-treated groups, pupillary miosis was not observed, the conjunctiva was not hyperemic, and there were no other outward manifestations of ocular inflammation (eg, blepharospasm, tearing, or other signs of discomfort).

There were no changes from baseline ophthalmic observations for animals given sterile saline, vehicle control article, or the isotype-matched human monoclonal antibody control (traztuzumab).

Studies examining the time course and resolution of the ocular inflammation following a single dose (study 6) showed that the anterior chamber inflammatory response was greatest 1 day after dosing and spontaneously and steadily improved over the 1-week follow-up period (Figure 3). The intensity and the time course of the inflammation were similar following an intravenous bolus injection and following a 2-hour intravenous infusion.

Immunolocalization of CNTO95 Binding to Normal Macaque and Human Eye Tissues

Table 2 summarizes the immunolocalization of CNTO95 to cryosections of normal macaque and human eyes. CNTO95 bound to various elements in both macaque and human eyes, and the staining patterns were similar in both macaques and humans.

In the eye, the staining was described as cytoplasmic staining, whereas in the positive control tissue, human placenta, both cytoplasmic and membrane staining was described.

Concentrations of CNTO95 in Serum and Vitreous

Table 3 summarizes the concentrations of CNTO95 measured in serum and vitreous at 24 hours after the fourth biweekly intravitreal or intravenous administration of CNTO95. Low levels of CNTO95 were detected in the serum following intravitreal injection of CNTO95. The concentrations in the serum following intravitreal administration were approximately 3000-fold lower than in the serum following intravenous administration. Following intravenous administration of CNTO95, 100-fold lower levels of CNTO95 were detected in the vitreous when compared with the intravitreal injection. The concentrations of CNTO95 in the vitreous following intravenous administration of CNTO95 were approximately 1000-lower than the serum concentrations.

Discussion

The observations from these studies show that CNTO95, a fully human α_νintegrin monoclonal antibody, induces ocular inflammation in macaques. This inflammation is transient and does not appear to be associated with any long-term toxicological consequences.

Transient anterior chamber inflammation is not an unexpected finding following intravitreal administration of a human protein because this observation has been reported for other human proteins, for example, a human anti-HER2 monoclonal antibody ¹⁹ and a human anti-VEGF antibody fragment (Lucentis; Lucentis approval information [e-mail]). ^16,17 Ocular inflammation following intravenous administration of a human monoclonal antibody is, however, an unexpected finding. The reason for the ocular inflammation following intravitreal or intravenous administration has not been established.

Published studies have shown that intravitreal administration of endotoxin (1–100 ng) can induce ocular inflammation consisting of anterior chamber inflammatory cells, aqueous flare, and corneal keratic precipitates in nonhuman primates. ^20,21 Also, systemic injection of endotoxin (100–200 μg) is a wellestablishedmeans of inducing exclusive ocular inflammation in rats. ²² Although endotoxin can induce an ocular inflammation similar to that described following administration of CNTO95, the concentration of endotoxin in the CNTO95 product was very low (<0.03 EU/mg). When administered intravitreally, a maximum of 0.015 EU of endotoxin was administered (0.03–0.75 ng, assuming 1 ng of endotoxin is approximately equivalent to 2–50 EU). When injected intravenously, a maximum of 4.5 EU (9–225 ng) of endotoxin was injected. The amount of endotoxin administered intravenously to the macaques is therefore considerably less than the amount administered systemically to rats to induce ocular inflammation. Also, the HER2 monoclonal antibody product contained a similar level of endotoxin and did not induce ocular inflammation. Therefore, endotoxin is unlikely to be the cause of the inflammation following intravenous administration of CNTO95. The inflammatory response is not a nonspecific reaction to administration of a human monoclonal antibody in macaques because the anti-HER2 mAb, which is also a human IgG1 antibody, did not induce ocular inflammation.

This transient inflammation did not appear to result in any long-term adverse consequence in the macaques because color ocular photographs and fluorescein angiography conducted after multipledose administration did not reveal any signs of ocular toxicity. In addition, histopathological examination of eyes from macaques treated intravenously with CNTO95 once per week at dosages up to 50 mg/kg for up to 6 months did not show any signs of ocular toxicity. However, in only one of the repeated-dose toxicology studies were ophthalmic examinations conducted following the first-dose administration of CNTO95 (study 8), and no findings were observed in that study. In subchronic and chronic repeateddose toxicology studies, it is not common practice to include ophthalmic examinations following the firstdose administration. The ophthalmic examinations are generally conducted only after multiple-dose administration. It is therefore not known whether the animals in the first 3 repeated-dose toxicology studies (studies 1–3) had developed a transient response that had resolved with repeated dosing or whether these animals did not develop the response. In all of the studies in which ophthalmic examinations were conducted within 1 week after the first dose and inflammatory cell and flare were identified, no outward signs of ocular toxicity were observed. Therefore, the ocular findings would likely go undetected if early ophthalmic examinations were not conducted.

Analysis of serum and vitreal samples for CNTO95 concentration demonstrated that a very small amount of CNTO95 was able to penetrate from the blood to the vitreous and vice versa. It should be emphasized that thesemeasurements were taken from eyes that had laser-induced burns and may therefore bemore “leaky” than normal undisturbed eyes. Therefore, this degree of penetration may overestimate the ocular exposure in normal eyes. However, even taking this caveat into consideration, the concentration of CNTO95 in the vitreous was one thousandth of the serum concentration following intravenous dosing. This low concentration of CNTO95 (∼2.5 μg/mL) in the eye is unlikely to produce any direct toxic effects on the eye. Studies conducted in animals and humans have suggested that concentrations of CNTO95 in excess of 10 μg/mL are required for saturation of α_ν-integrin binding. ^14,23 Immunohistochemical staining for CNTO95 binding to sections of eye from normal macaques and humans showed binding ofCNTO95 (10 μg/mL) to a number of ocular structures. However, the staining was described as cytoplasmic in contrast to the membrane staining observed in the positive control tissue (human placenta). Because antibodies are too large to diffuse across membranes, the cytoplasmic staining observed in macaque and human eyes is unlikely to be of clinical relevance. ^24,25 Therefore, immunohistochemical data also do not support a direct toxic effect of CNTO95 on the eye.

“Uveitic-like” reactions have also been observed in patients treated with CNTO95. In a phase I clinical study in cancer patients, ophthalmic examinations were initiated following the observations in the cynomolgus macaques. Ophthalmic examinations were conducted in the last 4 patients in that study, after they had received more than 1 infusion of CNTO95 (10 mg/kg), and no ophthalmologic findings were observed. ¹¹ However, in a phase II study in melanoma patients, in which ophthalmic examinations were included at the onset of the study, 19% of patients receiving 5 mg/kg of CNTO95 and 30% of patients receiving 10 mg/kg of CNTO95 experienced lowgrade, asymptomatic, uveitic-like reactions, which resolved without sequelae. ²⁴ The clinical observations were temporally similar to the observations in the macaques. In both the phase I and phase II clinical studies, the most frequent adverse events following intravenous administration of CNTO95 were infusion reactions consisting of nausea, fatigue, fever, and headache that occurred within 8 hours of CNTO95 administration and were dose dependent (32% at 5 mg/kg, 39% at 10 mg/kg in the phase II study). ²⁶ Similar acute infusion reactions have been reported in patients for a number of monoclonal antibody therapies. ^12,13 Consistent with the observations in the eye, the infusion reactions in the patients are greatest after the first-dose administration and diminish with subsequent administration. The mechanism of these first-dose infusion reactions has not been elucidated, but they can generally be prevented by pretreatment of patients with antihistamines, steroids, or acetaminophen.

The ocular inflammation that has been observed in patients and macaques may be a secondary manifestation of a first-dose infusion reaction even though macaques do not display any of the other signs of acute infusion reactions that are observed in patients.

In patients, the maximum serum concentration of CNTO95 following the first infusion at 10 mg/kg was 226 μg/mL. ¹¹ In macaques, the serum concentrations following the first infusions at 2, 10, and 50 mg/kg were 64, 296, and 1320 mg/kg, respectively. ¹⁴ In these studies, uveitic-like reactions were seen in macaques at all dose levels tested (2, 10, and 50 mg/ kg). Therefore, the uveitic-like reactions, as well as the infusion reactions, seen in the patients occur at serum concentrations that are similar to those that produce uveitic-like reactions in the macaques. In the toxicology studies in which CNTO95 was administered weekly at doses of 10 or 50 mg/kg, accumulation of CNTO95 occurred with steady state being obtained by week 7 and peak serum concentrations exceeding 2000 μg/mL. ¹⁴ Therefore, in the toxicology studies, the exposure of the macaques to CNTO95 at the end of the treatment period greatly exceeded the clinical exposure. In those studies, no ocular toxicity was observed at the end of the treatment periods.

In conclusion, treatment of cynomolgus macaques with CNTO95 either intravenously or intravitreally induced a transient ocular inflammation that may represent a secondary manifestation of a more generalized first-dose infusion reaction. This ocular inflammation did not appear to have any long-term ocular toxicological consequences.