Figure 17.1 Vascular endothelial growth factor (VEGF) is a homodimeric glycoprotein that is secreted in response to hypoxia and ischemia. VEGF induces angiogenesis and vascular permeability. Arrows show the binding site VEGFR. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
VEGF-A has been shown to cause neovascularization and leakage in models of ocular angiogenesis and is thought to contribute to the progression of AMD. Inhibition of inappropriate VEGF activity is a so-called antiangiogenic approach to the treatment of these diseases. The binding of ranibizumab to VEGF-A prevents the interaction of VEGF-A with its receptors (VEGFR1 and VEGFR2) on the surface of endothelial cells, reducing endothelial cell proliferation, vascular leakage, and new blood vessel formation. Ranibizumab, an anti–human VEGF, affinity-matured Fab, has been developed by Genentech as a therapeutic agent for treating ocular vascular disease by intravitreal (ITV) injection.
Drug Biochemistry and Formulation
Affinity-matured ranibizumab is produced by standard recombinant technology methods in an Escherichia coli expression vector and bacterial fermentation. Ranibizumab is not glycosylated and has a molecular mass of approximately 48,000 Da.
The lyophilized form of the drug is produced from a freeze-drying process. The solid cake seen in each vial consists of protein and the excipient components that are stabilizing agents for the protein. The lyophilized form of ranibizumab requires reconstitution with Sterile Water for Injection (SWI), USP. After reconstitution with SWI, the highly concentrated protein solution is further diluted with vehicle prior to ITV administration. This presentation was useful in early clinical dose-ranging studies and was used in early phase I/phase II studies.
The liquid formulation is the commercial form. This presentation provides excellent stability and eliminates the need for reconstitution and further dilution. Liquid ranibizumab is formulated as a sterile solution aseptically filled in a sterile, 2-mL glass vial. Each vial is designed to deliver 0.05 mL of 10 mg/mL of ranibizumab aqueous solution with 10 mM histidine HCI, 10% α,α-trehalose dehydrate, and 0.01% polysorbate 20 (pH 5.5). The vial contains no preservative and is suitable for single use only. Vials should be protected from light. The liquid formulation was assayed for purity and identity as described for the lyophilized form. The assay for potency is a biologically relevant, cell-based assay in which the ability to inhibit VEGF-dependent growth in vitro with a cell line requiring VEGF for growth is quantified.
VEGF THE TARGET
Human VEGF exists as at least six isoforms (VEGF121, VEGF145, VEGF165, VEGF183, VEGF189, and VEGF206) that evolve from alternative splicing of mRNA of a single gene (1). VEGF165, the most abundant isoform, is a basic, heparin binding, dimeric glycoprotein with a molecular mass of approximately 45,000 Da (1). Two VEGF receptor tyrosine kinases, VEGFR1 and VEGFR2, have been identified (2–7). VEGFR1 has the highest affinity for VEGF, with a Kd of approximately 10 to 20 pM (8), and VEGFR2 has a somewhat lower affinity for VEGF with a Kd of approximately 75 to 125 pM (9–11).
VEGF has numerous biologic functions, including regulation of VEGF gene expression under hypoxic conditions (1), mitogenic activity for micro- and macrovascular endothelial cells (4–7,12–14), and induction of expression of plasminogen activators and collagenase (15). Of particular significance to wet AMD are the angiogenic properties of VEGF, which have been demonstrated in a variety of in vivo models, including the chick chorioallantoic membrane (15,16), rabbit cornea (17), and rabbit bone (17). VEGF also functions as a survival factor for newly formed endothelial cells (16,18). Consistent with prosurvival activity, VEGF stimulates expression of the antiapoptotic proteins Bcl-2 and A1 in human endothelial cells (19). VEGF has been shown to induce vascular leakage in guinea pig skin (19). Dvorak et al. (20,21) suggested that an increase in microvascular permeability is a crucial step in angiogenesis associated with tumors and wound healing. According to this hypothesis, a core function of VEGF in the angiogenic process is the induction of plasma protein leakage. This outcome would result in the formation of an extravascular fibrin gel, which serves as a substrate for endothelial cells. This proposal may have substantial bearing for AMD as it is well known that permeability of the choroidal neovascularization (CNV) membrane results in the transduction of serum components beneath and into the retina, creating serous macular detachment, macular edema, and vision loss.
VEGF is expressed in an assortment of cells in the normal human retina. Colocalization of VEGF mRNA and protein is detected in the ganglion cell, inner nuclear and outer plexiform layers, the walls of the blood vessels, and photoreceptors (22). Retinal pigment epithelium (RPE), Müller cells, pericytes, vascular endothelium, and ganglion cells all manufacture VEGF (23,24).
Studies have documented the immunohistochemical localization of VEGF in surgically resected CNV membranes from AMD patients. Kvanta et al. (20) demonstrated the presence of VEGF mRNA and protein in RPE cells and fibroblast-like cells. Lopez et al. (19) noted that the RPE cells that were strongly immunoreactive for VEGF were present primarily in the highly vascularized regions of CNV membranes, whereas the RPE cells found in fibrotic regions of CNV membranes showed little VEGF reactivity. Kliffen et al. (25) also demonstrated increased VEGF expression in RPE cells and choroidal blood vessels in maculae from patients with wet AMD compared with controls.
An increase in VEGF expression has been noted in experimental models of CNV in rats and in nonhuman primates (26). In addition, transgenic mice with increased VEGF expression in photoreceptors (27) or RPE (28) developed neovascularization reminiscent of CNV seen in humans with neovascular AMD. This further supports the participation of VEGF in ocular neovascularization. These lines of evidence suggest that VEGF is a practical target for therapeutic intervention in neovascular AMD.
Inhibition of VEGF Activity by Ranibizumab
In Vitro Evaluation of Ranibizumab
A number of antibodies that can bind VEGF or inhibit VEGF activity were considered in vitro and in vivo for molecule selection. muMAb VEGF A.4.6.1 is a full-length murine monoclonal antibody of the IgG1 isotype that continuously and potently defuses the biologic activities of VEGF, including the endothelial cell mitogenic activity, vascular permeability–enhancing activity, and angiogenic properties in the chick chorioallantoic membrane (29). This antibody also inhibits growth of various human tumor types in animal models1, recognizes all isoforms of VEGF, binds to VEGF with a Kd of approximately 8 × 10−10 M, and neglects to recognize other peptide growth factors. Bevacizumab (rhuMAb VEGF) is a humanized version of muMAb VEGF A.4.6.1 that was produced by site-directed mutagenesis (30) (Figs. 17.2 and 17.3).
Figure 17.2 The humanization of muMAb VEGF A.4.6.1 involved the transfer of six complimentary-determining regions (CDR) from muMAb VEGF A.4.6.1 to a human framework by site-directed mutagenesis. The final antibody is a rhuMAb (recombinant, humanized, monoclonal antibody). (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
Figure 17.3 Histoautoradiograph of 125I-rhuMab VEGF Fab V1, the humanized Fab antibody (Column B), and 125Iodine (I)-rhuMab HER2, the full-length humanized antibody (Column A) following bilateral intravitreal injections in rhesus monkeys. The Fab fragment penetrated all of the layers of the retina evenly extending as far back as the RPE while the full-length antibody failed to penetrate farther than the ILM at any point during the study. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
Ranibizumab is a recombinant humanized antibody Fab fragment that neutralizes VEGF. Ranibizumab shows high affinity for binding VEGF (Fig. 17.4). In addition, ranibizumab potently inhibits survival of human umbilical vein endothelial cells (HUVEC) stimulated with 5 ng/mL of recombinant human VEGF. Ranibizumab also inhibits VEGF-mediated vascular permeability in the Miles assay (Fig. 17.5).
Figure 17.4 HUVEC proliferation assay. rhuFab V2 Lucentis is capable of binding all three VEGF isoforms and thereby inhibits VEGF isoforms–induced endothelial cell proliferation. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
Figure 17.5 RhuMAb VEGF was fragmented and affinity matured through complementarity-determining region (CDR) mutation and affinity selection by monovalent phage display. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
Nonclinical pharmacology studies to measure the safety of ITV administration of ranibizumab during and following laser-induced CNV in cynomolgus monkeys were conducted. Study 99-166-1757 evaluated single-agent ranibizumab treatment, while study 00-580-1757 assessed ranibizumab in combination with verteporfin photodynamic therapy (PDT). The findings of these studies point out that ITV administration of ranibizumab during laser induction of a choroidal neovascular membrane and during its subsequent formation do not exhibit any irregular toxicity. Moreover, treatment with ranibizumab in the cynomolgus monkey model prevents membrane formation and attenuates permeability from already-formed choroidal neovascular membranes (31) (Figs. 17.6 and 17.7).
Figure 17.6 The VEGF and the affinity-matured Fab fragment complex. Locally, it was possible to improve the contact between the antibody and the antigen through two mutations that improved the hydrogen bonding and van der Waals contact. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
FIGURE 17.7 A. Activation of the VEGF receptor (VEGFR1 and VEGFR2) of the endothelial cell by the VEGF generates an endothelial cell activation that increases the vascular permeability and induces angiogenesis. B. Lucentis binds to the receptor binding site of active forms of VEGF-A, avoid the interaction of VEGF-A with its receptors, on the surface of endothelial cells, reducing endothelial cell proliferation, vascular leakage, and new blood vessel formation. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
Following ITV administration in normal rabbits and cynomolgus monkeys, ranibizumab was cleared from the vitreous humor with a half-life of approximately 3 days and was distributed to all layers of the retina and cells where VEGF is expressed. In rabbits, a fraction of ranibizumab was cleared from the vitreous humor and through the choroidal capillaries. Concentrations of aqueous humor declined in parallel with those in the vitreous humor but were approximately 17 times lower. Serum ranibizumab concentrations were approximately 1/1,000 of vitreous concentrations and could not be quantified beyond Day 7 (after IV administration, the mean terminal half-life of ranibizumab was 3.1 hours). The systemic bioavailability of ranibizumab following ITV administration was estimated to be approximately 67%, suggesting that intraocular metabolism is not the predominant mechanism for clearance of ranibizumab from the vitreous humor. As expected with administration of a human protein into rabbits, antibodies to ranibizumab were detected in the vitreous humor by Day 7 but did not seem to affect the decline of ranibizumab concentrations in this compartment. Day 14 also detected antibodies to ranibizumab in serum. In cynomolgus monkeys, there did not appear to be dose-related differences for vitreous humor clearance and steady-state volume. Dose-normalized areas under the curves (AUCs) were also similar and suggested no dose-related differences in ranibizumab vitreous humor pharmacokinetics. In aqueous humor and retina, the half-life of ranibizumab was similar to that in vitreous humor, while the AUC was approximately two- to threefold lower. The apparent elimination half-life of ranibizumab in serum was similar to that in vitreous humor, while after IV administration, it was approximately 15 hours. Ranibizumab concentrations in serum increased proportionally with dose and were 400 to 1,500 fold lower than those in vitreous humor. The systemic bioavailability of ranibizumab after bilateral ITV administration of 0.5 mg and 2 mg per eye was estimated to be 60% and 50%, respectively. Regardless of dose, antibodies against ranibizumab were detected in the serum of some of the animals following the second dose; the incidence of antibodies tended to increase with increasing dose. Consistent with the estimates of half-life in vitreous and serum, ranibizumab did not appear to accumulate in the serum of antibody-negative animals when it was administered every 14 days in rabbits (Fig. 17.8).
Figure 17.8 Krzystolik assay. Cynomolgus monkeys, model of choroidal neovascularization. A. Pre-rhuFab injection. B. After rhuFab injection. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
Nonclinical single and multidose toxicology studies were conducted to support the clinical use of ranibizumab administered up to once every 2 weeks by ITV. Based on flare and cell responses, 0.5 mg ranibizumab per eye were considered to be the maximum tolerated dose for a single dose in cynomolgus monkeys. In summary, the available data have shown that ranibizumab is successful in suppressing CNV in nonhuman primates.
In summary, ITV administered ranibizumab was shown to diffuse to sites of VEGF expression in the retina of rabbits and monkeys. Ranibizumab was cleared relatively slowly, with a half-life of approximately 3 days from the vitreous humor of both species. Combined the in vitro data, the in vivo ocular pharmacokinetics in animals, and the extrapolation to humans (adapting for ocular size and using the mean pharmacokinetic parameters obtained in monkeys) strongly recommend that concentrations capable of inhibiting neovascularization in the retina can be attained in humans following ITV ranibizumab doses of 0.3 to 0.5 mg per eye.
Ranibizumab in Humans: Phase I/II Studies
Three studies were designed to investigate the doses of ranibizumab that would be appropriate for the treatment of neovascular AMD in large, multicenter, randomized phase I/II clinical trials. The first, FVF 1770g, a phase I dose-escalating study with patients with neovascular AMD (32); the second, a larger, phase I/II clinical study, investigated the tolerability and efficacy of multiple monthly ITV injections of ranibizumab at doses of 0.3 mg or 0.5 mg in patients with neovascular AMD (33); and the third trial, a phase I clinical study, investigated doses of ranibizumab above 0.5 mg to determine if doses higher than the maximum tolerated single dose could be well accepted when injected in an escalating stepwise fashion every 2 or 4 weeks (34). From the data analysis of these three clinical trials, the study team concludes that dose-limiting toxicity resulting in ocular inflammation was reached at 1,000 μg per eye. A dose of 500 μg per eye was determined to be the maximum tolerated dose; ITV injections of ranibizumab at escalating doses ranging from 0.3 to 2.0 mg were well tolerated; no serum antibodies against ranibizumab were observed (no antimouse immunoresponse that might decrease ranibizumab’s efficacy); the drug had a good safety profile with an improved VA and decreased leakage from CNV in subjects with neovascular AMD (Fig. 17.9).
Figure 17.9 Characterize the effect of rhuFab V2 (Lucentis) inhibition on VEGF-induced permeability in guinea pigs (Miles assay). (From Gaudreault J, Reich M, Arata A, et al. Ocular pharmacokinetics and antipermeability effect of rhuFab V2 in animals. Invest Ophthalmol Vis Sci. 2003;44(4):3942–3943.)
Based on these results, Genentech initiated another phase I/II study: the FOCUS Study. This trial was a single-masked, multicenter study evaluating the safety, tolerability, and efficacy of multiple-dose ITV injections of ranibizumab used in combinations with PDT in patients with predominantly classic AMD. After 2 years, the study showed a clear visual acuity benefit of adding ranibizumab to PDT treatment (35) (Fig. 17.10).
Figure 17.10 Mean change in visual acuity in FOCUS study. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)
The EXTENSION Study (FVF2508g) was a study to evaluate the long-term safety and tolerability of ranibizumab given as multiple ITV injections in subjects with primary or recurrent subfoveal CNV secondary to AMD who had completed the treatment phase of a Genentech-sponsored phase I or phase I/II ranibizumab study. The subjects were followed for treatment duration of up to 1,345 days, and subjects received up to 44 ITV injections. When the extension study concluded, at the time it represented the longest period of study of ranibizumab. The safety findings supported the further study of ranibizumab for long-term administration in subjects with neovascular AMD. Repeated ITV injection of ranibizumab was associated with a low rate of serious ocular adverse events, and it commonly induced mild intraocular inflammation. Ranibizumab was well tolerated systemically after multiple ITV injections. There was a low incidence of serum immunoreactivity to the study drug. Positive serum immunoreactivity did not appear to be associated with adverse events such as intraocular inflammation or marked decrease in visual acuity; however, the number of patients with positive antitherapeutic antibodies was small. Although the study was not designed to assess efficacy in terms of visual acuity, most subjects treated in the extension study experienced improved or preserved visual function.
There have been several phase III/IV clinical trials developed by Genentech in the United States during the past 9 years to prove the efficacy and safety of ranibizumab: MARINA, ANCHOR, HARBOR, PIER, SAILOR, and PrONTO.
MARINA Study (FVF 2598g)
The MARINA study was the first phase III clinical trial designed. It was a multicenter, randomized, double-masked, sham injection–controlled study of the efficacy and safety of ranibizumab in subjects with minimally classic or occult subfoveal neovascular AMD (36). Subjects were randomized in a 1:1:1 ratio to receive 0.5-mg ranibizumab, 0.3-mg ranibizumab, or a sham injection administered monthly (30 ± 7 days) for up to a maximum of 24 injections during the 2-year study period. Approximately 3 months prior to the study’s completion, subjects in the sham injection group who still remained on treatment were offered the opportunity to cross over to receive 0.5-mg ranibizumab for the remainder of the treatment period. The study met its primary end point, with nearly 95% of ranibizumab-treated subjects maintaining or improving vision at 12 months, compared with 62% of sham-treated subjects. The benefit of ranibizumab treatment over sham injections increased further through 24 months (Fig. 17.11). Ranibizumab administered as monthly ITV injections of 0.3 mg or 0.5 mg over 24 months was safe and well tolerated by subjects with minimally classic or occult subfoveal neovascular AMD.
Figure 17.11 Mean change in visual acuity in MARINA study. (Reprinted with permission from Quiroz-Mercado H, Kerrison JB, Alfaro DV, et al. Macular surgery, 2nd ed. Philadelphia, PA: Lippincott Williams & Wilkins; 2011.)