Lampalizumab is an antigen-binding fragment of a humanized monoclonal antibody against complement factor D (CFD), a rate-limiting enzyme in the activation and amplification of the alternative complement pathway (ACP), which is in phase III clinical trials for the treatment of geographic atrophy. Understanding of the pharmacokinetics, pharmacodynamics, and biodistribution of lampalizumab following intravitreal administration in the ocular compartments and systemic circulation is limited but crucial for selecting doses that provide optimal efficacy and safety. Here, we sought to construct a semimechanistic and integrated ocular-systemic pharmacokinetic-pharmacodynamic model of lampalizumab in the cynomolgus monkey to provide a quantitative understanding of the ocular and systemic disposition of lampalizumab and CFD inhibition. The model takes into account target-mediated drug disposition, target turnover, and drug distribution across ocular tissues and systemic circulation. Following intravitreal administration, lampalizumab achieves rapid equilibration across ocular tissues. Lampalizumab ocular elimination is relatively slow, with a τ1/2 of approximately 3 days, whereas systemic elimination is rapid, with a τ1/2 of 0.8 hours. Target-independent linear clearance is predominant in the eye, whereas target-mediated clearance is predominant in the systemic circulation. Systemic CFD synthesis was estimated to be high (7.8 mg/day); however, the amount of CFD entering the eye due to influx from the systemic circulation was small (<10%) compared with the lampalizumab dose and is thus expected to have an insignificant impact on the clinical dose-regimen decision. Our findings support the clinical use of intravitreal lampalizumab to achieve significant ocular ACP inhibition while maintaining low systemic exposure and minimal systemic ACP inhibition.
Geographic atrophy (GA) is an advanced stage of age-related macular degeneration (AMD) and is characterized by the loss of retinal photoreceptors, retinal pigment epithelium (RPE), and choriocapillaris. Vision loss in GA is progressive, and approximately 26% of legal cases of blindness in United Kingdom are due to GA (Rees et al., 2014; United Kingdom Department of Health, 2013). Worldwide, there are more than 5 million people with GA (Sunness et al., 1997; Wong et al., 2014). Despite the increasing prevalence of GA, it currently remains a largely unmet clinical need as no approved therapy is available (Girmens et al., 2012). The pathogenesis of AMD is not well understood; however, genetic and environmental factors appear to contribute to the disease process (de Jong, 2006). The alternative complement pathway has been implicated in AMD by both human genetics and histopathology. Genetic epidemiology studies have demonstrated that polymorphisms in alternative complement pathway–associated genes have strong correlations with the risk of developing AMD (Fritsche et al., 2013). In addition, activated components of the alternative complement pathway have been found in drusen (lipoproteinous depositions in the space between the RPE and Bruch’s membrane), which are a hallmark of AMD. These findings suggest that the alternative complement pathway may be an important mediator of AMD.
Lampalizumab (Genentech, Inc., South San Francisco, CA) is the antigen-binding fragment (Fab) of a humanized monoclonal antibody (mAb) that inhibits complement factor D (CFD), which is the rate-limiting enzyme in the activation and amplification of the alternative complement pathway (Katschke et al., 2012). Lampalizumab intravitreal (ITV) injection is under evaluation for the treatment of GA. The phase Ia (NCT00973011) (Do et al., 2014) and the MAHALO phase Ib/II clinical trials (NCT01229215) have been completed. Lampalizumab was shown to have an acceptable safety profile when administered at 10 mg per eye (Do et al., 2014). Two phase III, double-masked, multicenter, randomized, sham injection-controlled studies [NCT02247479 (Chroma) and NCT02247531 (Spectri)] are under way.
Retinal diseases, such as AMD, are difficult to treat, partly because the anatomy of the eye makes drug delivery to its posterior segment challenging (Kang-Mieler et al., 2014). Intravitreal delivery is the most direct drug delivery method to treat retinal diseases. Understanding of the pharmacokinetics (PK), pharmacodynamics (PD), and biodistribution of such an ITV-administered drug and its target in the ocular compartments and systemic circulation is limited but crucial to allow for optimal dosing and ensure efficacy and safety. PK/PD modeling provides a useful quantitative framework to evaluate ocular and systemic drug disposition as it accounts for all processes governing the PK characteristics (such as antigen binding, distribution, and clearance processes) and corresponding PD effects (Agoram et al., 2007; Van Der Graaf and Gabrielsson, 2009; Luu et al., 2013; Schuck et al., 2015). A target-mediated drug disposition (TMDD) model is critical in explaining the PK and PD of many biologics for which the high antibody-target binding affinity as well as the antibody and target levels can directly influence the PK of the antibody (Mager and Jusko, 2001). Lampalizumab was shown to bind CFD in both human and cynomolgus monkey (Macaca fascicularis) species with high affinity (Loyet et al., 2014). Cynomolgus monkeys were shown to be a relevant preclinical model for evaluating the PK/PD of lampalizumab (Katschke et al., 2012; Loyet et al., 2014). A single ITV dose-escalation study, a 2-week single-dose ITV toxicity study, and a PK study of i.v. or ITV injections of lampalizumab were performed in cynomolgus monkeys. The objective of this analysis is to develop a semimechanistic and integrated ocular-systemic PK/PD model of lampalizumab in a cynomolgus monkey to provide a quantitative understanding of the disposition of lampalizumab and inhibition of CFD in both ocular tissues and systemic circulation following ITV administration of lampalizumab.
Materials and Methods
Lampalizumab Pharmacokinetic and Pharmacodynamic Studies
Table 1 summarizes the designs of the three studies included in this analysis. These preclinical studies were conducted in both male and female cynomolgus monkeys at Covance (Madison, WI). The studies were approved by the Covance Institutional Animal Care and Use Committee. The studies were conducted in compliance with United States Food and Drug Administration regulations for Good Laboratory Practice for Nonclinical Laboratory Studies (Code of Federal Regulations Title 21, Part 58) and the U.S. Department of Agriculture Animal Welfare Act Regulations (Code of Federal Regulations Title 9, Part 3) Guide for the Care and Use of Laboratory Animals. Lampalizumab was administered via a single i.v. or ITV injection at the indicated dosage. For i.v. administration, lampalizumab was administered as a single bolus dose into the saphenous vein. For ITV administration, lampalizumab was administered as a single injection into both eyes.
Baseline (predose) and postdose blood samples for PK analysis were collected from each animal via a femoral vein at the time points indicated in Table 1. Blood samples were collected in serum separator tubes without an anticoagulant, allowed to clot at an ambient temperature for at least 20 minutes, and then centrifuged in a refrigerated centrifuge set to 2–8°C. The serum was harvested within 20 minutes of centrifugation and stored between −60°C and −80°C before PK analysis. Following euthanasia, vitreous humor, aqueous humor, and retinal tissue were collected from both eyes. The retinal tissue collected included neuroretina and RPE/choroid and excluded sclera. Additionally, a strip of retinal tissue, including the disc, macula, and peripheral retina, was collected. Retinal tissue samples were weighed and recorded. All ocular matrix samples were stored at −80°C before analysis.
Analysis of Lampalizumab and CFD
Lampalizumab in Cynomolgus Monkey Serum.
The concentration of lampalizumab in male and female cynomolgus monkey serum was measured by enzyme-linked immunosorbent assay (ELISA). A mouse mAb to lampalizumab (clone 7470; Genentech, Inc.) was adsorbed to the surface of microtiter plates overnight at 2–8°C to capture total lampalizumab (bound and unbound to CFD). After a 2-hour incubation, goat anti-human IgG–horseradish peroxidase (Bethyl, Montgomery, TX), which detects lampalizumab in the presence or absence of CFD, was added for 1 hour. The tetramethylbenzidine substrate (40 mM tetramethylbenzidine and 8 mM tetrabutylammonium borohydride in N,N′-dimethylacetamide) was subsequently added, and the reaction was stopped with 1 M phosphoric acid after optimal color development. Absorbance was measured photometrically at 450 nm and referenced to 650 nm. The standard curve ranged from 2.5 to 320 ng/ml. The minimum quantifiable concentration of total lampalizumab in cynomolgus monkey serum was 10 ng/ml.
Lampalizumab in Cynomolgus Monkey Aqueous Humor, Vitreous Humor, and Retinal Tissue.
The concentration of lampalizumab in cynomolgus monkey aqueous humor, vitreous humor, and retinal tissue was measured by ELISA as described above. The standard curve ranged from 12.5 to 1600 ng/ml. Retinal tissue samples were homogenized, and the homogenate was analyzed in the ELISA. Retinal tissue results were normalized by tissue weight. The minimum quantifiable concentration of total lampalizumab in cynomolgus monkey aqueous humor, vitreous humor, and retinal tissue was 26.3 ng/ml.
Factor D in Cynomolgus Monkey Serum.
The concentration of CFD in cynomolgus monkey serum was measured using ELISA. A mouse mAb specific to CFD (clone 4676; Genentech, Inc.) was adsorbed to the surface of microtiter plates overnight at 2–8°C to capture total CFD (bound and unbound to lampalizumab). Lampalizumab was then added to samples to saturate CFD-binding sites. After a 2-hour incubation, biotin-labeled mAb 7470 was added for 1 hour. Streptavidin–horseradish peroxidase (Thermo Scientific, Waltham, MA) was added for 1 hour. Substrate incubation and reading of the plate were performed as described for the lampalizumab serum method. The standard curve ranged from 3.9 to 500 ng/ml. The minimum quantifiable concentration of total CFD in cynomolgus monkey serum was 5 ng/ml.
Factor D in Cynomolgus Monkey Aqueous and Vitreous Humors.
The concentration of CFD in cynomolgus monkey aqueous and vitreous humors was measured by ELISA as described for the CFD serum method with the following modifications. The standard curve ranged from 0.8 to 100 ng/ml. The minimum quantifiable concentration of total CFD in cynomolgus monkey aqueous and vitreous humors was 2 ng/ml.
Target-Mediated Drug Disposition Model
All ocular and serum (following i.v. and ITV administration) data were simultaneously analyzed using a TMDD model that incorporates the binding equilibrium and in vivo turnover rates of free CFD, lampalizumab, and the drug-target (lampalizumab-CFD) complex as depicted by Fig. 1. NONMEM software (version 7.2; ICON Development Solutions, Dublin, Ireland) was used to perform all model estimation and simulation.
A quasi–steady state approximation (Gibiansky et al., 2008) of the TMDD model described the total vitreous lampalizumab concentrations (unbound and bound to CFD) and total vitreous CFD concentrations (unbound and bound to lampalizumab). The equations that describe the final model are shown below:(1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)Here, eqs. 1–4 describe the kinetics of the total lampalizumab amount in vitreous humor (AVITR), total CFD concentration in vitreous humor (RVITR), total lampalizumab amount in serum (ASER), and total CFD concentration in serum (RSER), respectively. Equations 5 and 6 describe the kinetics of the peripheral compartments for lampalizumab and CFD, respectively. Equations 7 and 8 describe the quasi–steady state approximation for the relationship between unbound lampalizumab concentrations in vitreous humor and serum (Cvu and Csu), total lampalizumab concentrations in vitreous humor and serum (CVITR and CSER), and total CFD concentration in vitreous humor and serum (RVITR and RSER), respectively. Equation 9 describes the relationship between aqueous (CAQ_tot) and vitreous (CVITR) total lampalizumab concentrations and the vitreous-aqueous lampalizumab partition coefficient (λAQ). Equation 10 describes the relationship between total CFD levels in aqueous humor (RAQ), vitreous humor (RVITR), and the vitreous-aqueous partition coefficient of total CFD (λTAQ). Equation 11 describes the relationship between retina (CRET_tot) and vitreous (CVITR) total lampalizumab concentrations and the vitreous-retina lampalizumab partition coefficient (λVR). The actual transfer rates for the complex, free drug, and free target may be different. However, as there are not sufficient data to build a more complex model, we simplified the peripheral distribution process by using only two rate constants (k12 and k21) for the drug rates and one rate constant (k12T = k21T) for the target rate. k12T was assumed to be the same as k21T. These are empirical parameters to improve the model goodness of fit of the i.v. arms.
VVITR and VC are the vitreous and serum volumes, respectively; RAQ is the total CFD concentration in the aqueous humor; kout, koutT, and koutC denote the ocular elimination rate constants of the unbound lampalizumab, free CFD, and lampalizumab-CFD complex, respectively; koutT was fixed to a value of 0.3 on day 1 based on the vitreous half-life of radioactively labeled factor D following ITV administration (unpublished in-house data). k, kdeg, and kinT denote the systemic elimination rate constants of the unbound lampalizumab, free CFD, and lampalizumab-CFD complex, respectively; ksyn is the zero-order production rate constant of CFD in the vitreous humor; k is the first-order clearance rate constant of lampalizumab in the serum; kinC is the first-order rate constant of the lampalizumab-CFD complex from the serum to the vitreous humor and was assumed to be negligible because the concentration of the complex is higher in the ocular environment following ITV administration; kSV is the first-order rate constant of free CFD from the serum to the vitreous humor; and KSS is the quasi–steady state constant of lampalizumab binding to CFD. KSS was fixed to the in vitro parameter of 11.7 pM (Loyet et al., 2014). F1 denotes the fraction of drug that enters the vitreous directly following ITV administration, and the remaining fraction (1-F1) was assumed to leak to the serum by fast absorption, presumably from the microcirculation.
A naïve-pool approach was used to model the data. The residual error model for the observations was best described by a proportional error model, and the proportional residual errors were assumed to be independent and normally distributed with zero means.
A series of model simulations were performed using the estimated model parameters to examine the behavior of the drug and target in ocular tissue and serum as a function of time and dose level to provide a better understanding of the PK/PD of lampalizumab. The fractions of the drug eliminated via linear clearance or target-mediated clearance at time t are defined by eqs. 14 and 15, respectively, in the eye, and eqs. 16 and 17, respectively, in the serum. The time course of total target flux from systemic to ocular (Rin) was simulated (eq. 18) to understand the relative total target influx from systemic circulation to ocular tissues. All simulations were performed using NONMEM software.(14)(15)(16)(17)(18)
Ocular and Systemic Lampalizumab and CFD Concentration-Time Profiles Are Well Described by a TMDD Model.
All available lampalizumab and CFD data (Table 1) in the ocular tissues and serum were combined and fit using a TMDD model (Fig. 1) with a single set of parameters. The observed concentration-time profiles of lampalizumab and total CFD in the serum following i.v. and ITV administration were well described by the proposed combined ocular-serum PK/PD model. The model-predicted time courses of both lampalizumab and total CFD were in good agreement with the observed data at different dose levels in both the serum (Fig. 2, A and B) and ocular (Fig. 2, C–E) compartments. All model parameters were estimated with good precision, with relative standard errors (RSEs) below 24%, except for the CFD distribution rate constant (53.8%) (Table 2).
The estimated ocular elimination half-life of lampalizumab was 2.9 days (kout = 0.2 day−1; RSE = 2.1%) and was markedly longer than the systemic elimination half-life of 0.8 hours (k = 21.3 day−1; RSE = 6.6%). The volumes of distribution of lampalizumab in the vitreous humor and serum were 2.2 and 127 ml (∼32–64 ml/kg for 2–4 kg body weight in this study), respectively. These values are in broad agreement with the literature values for vitreous humor volume (Struble et al., 2014) and antibody volume distribution (Deng et al., 2011) in the cynomolgus monkey. For CFD, the elimination half-life from the serum was estimated to be 0.2 hours (kdeg = 96.0 day−1; RSE = 10.1%) and the synthesis rate was 2.6 nmol/ml per day or 7.8 mg/day. The elimination half-lives of the lampalizumab-CFD complex from the vitreous humor and serum were 0.9 days (koutC = 0.7 day−1; RSE = 16.9%) and 0.2 days (kint = 4.2 day−1; RSE = 5.6%), respectively.
Distribution across Ocular Tissues Achieved Quick Equilibrium.
Lampalizumab concentrations were determined in vitreous humor, aqueous humor, and the retina to help elucidate the pharmacokinetics within ocular tissues. Elimination rates of lampalizumab in the vitreous humor, aqueous humor, and retina were similar (half-life = 2.9 days). Maximum tissue concentrations in the vitreous humor, aqueous humor, and retina were observed at the first observed data point of 10 hours, which was shorter than the ocular elimination half-life of 2.9 days, suggesting that ocular tissue distribution achieved equilibrium relatively quickly. This was further supported by evidence that the aqueous humor, vitreous humor, and retina exposures can be related by a constant partition coefficient. The lampalizumab vitreous-aqueous partition coefficient was estimated to be 4.4 (RSE = 12.1%), and the vitreous-retina partition coefficient was 7.3 (RSE = 15.3%). Similarly, the partition coefficient of CFD between the vitreous and aqueous humors was 0.5 (RSE = 23.5%).
Lampalizumab Administration Achieved High Ocular Target Occupancy with Transient Systemic CFD Inhibition.
Ocular tissue concentrations of both lampalizumab and total CFD were determined to elucidate ocular pharmacokinetics and target occupancy with ITV administration of lampalizumab. Ocular lampalizumab levels in both the vitreous and aqueous humors were well above the total CFD level (at least 10× higher) over the 16 days of the study (Fig. 2, C and D). Simulation of 1–10 mg of lampalizumab per eye showed that the vitreous total drug level exists as a predominantly free drug and that the total target level exists as a predominantly drug-target complex for the first 30–40 days following dosing (Fig. 3A). The free target level is predicted to be well below (3–6 orders of magnitude below) the baseline target level over this period of time. In addition, ocular lampalizumab clearance was predominantly through the linear clearance mechanism (more than ∼99% of the total clearance), with minimal target-mediated clearance over a simulated dose range of 0.1–50 mg/eye (Fig. 3B).
The time courses of systemic lampalizumab and total CFD concentrations were quantified to assess the systemic pharmacokinetics of lampalizumab following ITV administration and potential systemic inhibition of CFD.
The systemic pharmacokinetics of lampalizumab exhibited flip-flop kinetics, a phenomenon in which the systemic lampalizumab pharmacokinetics were limited by the ocular-to-serum absorption rate, and the serum terminal elimination rate in this case was reflective of the ocular-to-serum absorption rate (Fig. 2, A and B). In addition, systemic elimination following i.v. administration (k = 21.3 day−1) is substantially faster than that following ITV administration (kout = 0.2 day−1) (Table 2). As a result of this rapid systemic clearance, the systemic lampalizumab level in the serum was low, with concentrations below the total CFD level for most of the 24 days of the study (Fig. 2B). Simulation of 1–10 mg of lampalizumab per eye showed that the serum total drug level existed as a predominantly drug-CFD complex for up to 60 days following dosing (Fig. 4A). The free target level only dropped transiently, with nadir at 2 and 19% below baseline for 1 and 10 mg, respectively, and recovered back to the baseline target level within 10 days following dosing. In addition, systemic lampalizumab clearance was predominantly through the target-mediated clearance mechanism (more than 90% of total clearance) over a large dose range (0.1–50 mg/eye) (Fig. 4B).
Target occupancy in the vitreous humor and serum was simulated using the TMDD model. The free CFD level in the vitreous humor dropped substantially upon ITV lampalizumab dosing, and maximum vitreous CFD inhibition occurred quickly (within 1 day following ITV dosing). More than 95% target occupancy was maintained for 34 and 44 days following a single ITV dose of 1 and 10 mg/eye, respectively (Fig. 5). Systemic CFD inhibition was transient, with target occupancy in the serum reaching maximum target occupancy at 41 and 89% for 1 and 10 mg per eye, respectively, within day 1 and decreasing approximately linearly over time. Recovery back to ∼10% occupancy was at 8 and 17 days after dosing (Fig. 5). Despite the moderate systemic target occupancy, the serum lampalizumab molar concentrations for the 1- or 10-mg dose groups either did not or only transiently exceeded the molar concentrations of CFD. This molar excess of lampalizumab was shown to be necessary for systemic alternative complement pathway inhibition based on the alternative complement pathway hemolytic activity assay (Loyet et al., 2014).
Ocular Influx of CFD Following Intravitreal Lampalizumab Dosing.
To study the extent of systemic-to-ocular influx of CFD, we examined the cumulative ocular influx over time as a function of dose using model simulations. Despite the high CFD synthesis in the serum (2.6 nmol/ml per day), the serum-to-ocular transfer rate constant of CFD was estimated to be small (3×10−4 day−1) (Table 2). As a result, the cumulative ocular influx of CFD at 30 days following ITV dosing was small compared with the total administered drug amount at less than 10% and less than 1% of the 1- and 10-mg dose per eye, respectively (Fig. 6).
Target-mediated drug disposition, whereby the pharmacokinetics of a drug are significantly altered by the process of binding to its biologic target, was shown to be an important mechanism affecting the PK and PD of many biologics (Mager and Jusko, 2001; Mager, 2006). Understanding the TMDD of an investigative drug provides the understanding of drug levels and target occupancy at various tissue sites and is ultimately an important tool in selecting the right dose level and regimen for therapeutic biologics. This is especially important for ocular-targeted biologics, for which the target levels and turnover rates in the eye and systemic circulation can be very different, such as those for CFD, which has a systemic synthesis rate estimated at 1.33 mg/kg per day compared with relatively few neural retina CFD transcripts and an average systemic concentration of 42 nM, which is ∼2.5 times higher than in postmortem human vitreous humor (Barnum et al., 1984; Pascual et al., 1988; Anderson et al., 2010; Loyet et al., 2012). In this analysis, we sought to construct a semimechanistic PK/PD model of lampalizumab in both ocular tissues and serum in healthy cynomolgus monkeys. The intensive and invasive tissue sampling in the monkey model enables a mechanistic and integrated understanding of the interplay of the ocular and systemic PK/PD of lampalizumab. The results from this analysis provide a better understanding of the lampalizumab PK/PD and the constructed PK/PD model provides an important tool to predict and optimize future studies of lampalizumab for the treatment of geographic atrophy. For the purpose of extrapolating to humans, it should be noted that previous studies have shown plasma, but not vitreous humor, levels of CFD that were significantly elevated in AMD patients compared with control cases (Scholl et al., 2008; Hecker et al., 2010; Stanton et al., 2011; Loyet et al., 2012).
The serum and ocular lampalizumab and CFD concentration time courses following a large range of both i.v. and ITV doses (Table 1) were well described (Fig. 2) by a TMDD model that takes into account target binding, distribution, vitreous-aqueous and vitreous-retina partitioning, target turnover, and diffusion-driven transport processes across different ocular and serum compartments (Fig. 1). Following ITV administration, lampalizumab is distributed to the aqueous and retina relatively quickly, within the first 45 minutes following dosing. Lampalizumab partitions relatively similarly between the vitreous and aqueous humors (λVR = 4.4) and between the vitreous humor and retina (λVR = 7.3). The ocular elimination half-life of lampalizumab in cynomolgus monkeys was estimated to be approximately 3 days and was significantly longer than the systemic elimination half-life of 0.8 hours. A similar difference between the ocular and systemic half-life was noted in a TMDD model of lampalizumab in humans (Le et al., 2015), in which the vitreous half-life was 5.9 days compared with a systemic half-life of 9 hours. The shorter ocular half-life in monkeys (3 days) compared with humans (7–9 days) was also reported for ranibizumab (Gaudreault et al., 2005; Krohne et al., 2012; Xu et al., 2013). Similar to the clinical results, lampalizumab in the cynomolgus monkey also exhibits flip-flop pharmacokinetics, a phenomenon in which the rate of drug egressing out of the vitreous humor is much slower than the rate of drug cleared from the systemic circulation. This phenomenon is also similar to that exhibited by ITV ranibizumab, a Fab (Xu et al., 2013), and ITV pegaptanib, a pegylated aptamer (Macugen, 2011), and helps to maintain a low systemic exposure and minimize any potential systemic safety issues. There has been limited literature on the comparison of ocular PK/PD properties for full-length antibodies versus Fabs. However, our results suggested that there appears to be no gain in half-life with a higher molecular weight when comparing the ocular half-life of free lampalizumab, free CFD, and the lampalizumab-CFD complex (molecular weight ranges from 24 to 72 kD). The relationship between the ocular half-life and molecular size is unknown beyond this range. However, literature on ITV administration of anti–vascular endothelial growth factor therapy suggested that the half-life of ranibizumab (7–9 days) (Krohne et al., 2012; Xu et al., 2013) was only slightly shorter than that for bevacizumab (9.8 days) (Krohne et al., 2012) in humans. Notably, the quasi–steady state equilibrium constant Kss was fixed to the in vitro measured value of the dissociation equilibrium constant (KD) of 11.7 pM (Loyet et al., 2014) and was able to describe the data quite well. As the quasi-equilibrium approximation is a special case of the quasi–steady state approximation (Gibiansky et al., 2008), and the in vitro and in vivo agreement of the parameters KD and Kss confirms the assumption that the internalization rate constant is much smaller than the dissociation rate constant.
In the cynomolgus monkeys, ITV lampalizumab achieves near-complete target occupancy in the vitreous humor (Fig. 5A). This is evidenced by the dramatically higher observed ocular lampalizumab concentrations compared with total ocular CFD levels (Fig. 2, C and D), and the many orders of magnitude drop in simulated free CFD lasting more than 30 days following dosing (Fig. 3A). At the dose range of 1–10 mg per eye, the exposure ranges of lampalizumab in the vitreous humor, aqueous humor, and retina are above 10 nM (Fig. 2, C–E) during the 16-day study time and are well above the Kss and KD value of 11.7 pM (Loyet et al., 2014), suggesting that the maximal target-mediated clearance was reached and target-independent linear kinetics are predominant. In fact, Fig. 3B illustrates that over a large range of ITV dose levels (0.1–50 mg per eye), linear clearance is expected to be predominant. The linear clearance process is governed mainly by the rate of lampalizumab exiting the eye, as evidenced by the similar slopes in the ocular (aqueous, vitreous, and retina) concentrations and serum concentrations versus time (Fig. 2, B–E).
Upon exiting the eye, lampalizumab is cleared by two different mechanisms: a linear target-independent clearance mechanism and a target-mediated clearance mechanism. The linear clearance mechanism is characterized by a first-order systemic elimination half-life of 0.8 hours (Table 2). This parameter was well estimated (RSE = 1.2%) due to the availability of the serum lampalizumab and CFD data following i.v. administration. However, the linear clearance mechanism is only a minor pathway for serum lampalizumab clearance, which is composed of approximately 2–11% of the total clearance over the range of 0.1–50 mg lampalizumab per eye (Fig. 4B). The target-mediated clearance mechanism is responsible for approximately 90% of lampalizumab clearance in the serum as a result of a lower lampalizumab exposure and higher CFD level in the systemic circulation. The higher serum CFD concentrations compared with lampalizumab over time in the dosing groups in which both drug and target data were available suggest that target-mediated clearance is predominant. This is especially true due to the high binding affinity between lampalizumab and CFD (Loyet et al., 2014). The clear indication of target-mediated clearance is also shown by the increase in the total CFD concentrations upon both i.v. and ITV lampalizumab administration (Fig. 2, A and B). This also confirms the binding of lampalizumab to CFD in vivo as the lampalizumab-CFD complex is expected to clear more slowly than CFD alone, partly because of the larger size of the complex. This is in agreement with the estimated elimination constant parameters of CFD and the lampalizumab-CFD complex, 96 and 4.2 day−1, respectively (Table 2). Many studies in the literature also suggested similar mechanisms of slower clearance of the complex compared with the soluble target (Hayashi et al., 2007; Davda and Hansen, 2010).
Our analysis also revealed that the synthesis rate of CFD in the systemic circulation is 7.8 mg/day (or ∼2.6 mg/kg per day for a 3-kg monkey). This is similar to the synthesis rate of CFD in humans. The current literature suggests that rapid synthesis of CFD occurs in the systemic circulation, with an estimated rate of 1.33 mg/kg per day in humans (with a typical body weight of 75 kg) (Pascual et al., 1988). Despite the rapid synthesis of CFD, our simulation studies in the cynomolgus monkey model suggested that only a relatively small amount enters the eye. The cumulative ocular influx of CFD remains less than 0.1 molar fraction of the total dose at up to 30 days following ITV administration (Fig. 6). In a TMDD model in humans, the composite influx rate of CFD into the ocular compartment estimated from the study model was 0.32 µg/ml per day (Le et al., 2015). This accounts for less than 1% of the systemic production rate of 1.33 mg/kg per day.
In summary, our model-based PK/PD analysis of lampalizumab in cynomolgus monkeys represents the first analysis to mechanistically examine the interplay between ocular and systemic PK of an antibody in ophthalmology drug development. This study shows that lampalizumab distributes rapidly across the ocular tissues upon ITV administration. Lampalizumab PK exhibits flip-flop kinetics, with slow ocular elimination and rapid systemic elimination. Our analysis also suggests that the linear clearance mechanism is predominant in the ocular environment as a result of excess lampalizumab compared with CFD concentration, whereas a target-mediated clearance mechanism is predominant in the systemic circulation due to the low systemic exposure of lampalizumab following ITV administration in the face of high CFD levels resulting from high synthesis. Intravitreal administration of lampalizumab enables a high local lampalizumab concentration at the site of action while minimizing systemic lampalizumab exposure to ensure patient safety. Near-complete target occupancy is maintained over more than 1 month in the vitreous humor within the tested dose range, while transient target inhibition is observed in the systemic circulation following lampalizumab ITV administration, with PK profiles favorable for ocular efficacy and systemic safety. The systemic synthesis rate of CFD is rapid and consistent with the values in humans, but ocular influx of factor D into the eye is minimal compared with the total ITV drug dose and expected to have no impact on the dosing strategy. This study also demonstrated how the ocular microenvironment differs substantially from the systemic circulation with regards to the synthesis and turnover rates of the therapeutic target CFD. The ocular PK/PD parameters from our model may be helpful in predicting human ocular PK and PD for lampalizumab as it is often not possible to sample human ocular tissues.
Support for third-party writing assistance for this manuscript, furnished by Emma A. Platt, of Envision Scientific Solutions, was provided by Genentech, Inc.
Participated in research design: van Lookeren Campagne, Damico-Beyer.
Conducted experiments: Le, Gibiansky, Good, Davancaze, Loyet, Morimoto.
Contributed new reagents or analytic tools: Le, Gibiansky, Good, Davancaze, Loyet, Morimoto.
Performed data analysis: Le, Gibiansky.
Wrote or contributed to the writing of the manuscript: Le, Gibiansky, Good, Davancaze, van Lookeren Campagne, Loyet, Morimoto, Jin, Damico-Beyer, Hanley.
- Received July 7, 2015.
- Accepted September 9, 2015.
Genentech, Inc., South San Francisco, CA, provided support for the study and participated in the study design; conducting the study; and data collection, management and interpretation.
- age-related macular degeneration
- complement factor D
- enzyme-linked immunosorbent assay
- antigen-binding fragment
- geographic atrophy
- monoclonal antibody
- retinal pigment epithelium
- relative standard error
- target-mediated drug disposition
- Copyright © 2015 by The American Society for Pharmacology and Experimental Therapeutics