October 2015
Volume 56, Issue 11
Free
Physiology and Pharmacology  |   October 2015
Evaluation of Fluorophotometry to Assess the Vitreal Pharmacokinetics of Protein Therapeutics
Author Affiliations & Notes
  • Leslie J. Dickmann
    Department of Preclinical and Translational Pharmacokinetics Genentech, Inc., South San Francisco, California, United States
  • Victor Yip
    Department of Preclinical and Translational Pharmacokinetics Genentech, Inc., South San Francisco, California, United States
  • Cuiwen Li
    Department of Preclinical and Translational Pharmacokinetics Genentech, Inc., South San Francisco, California, United States
  • Juan Abundes, Jr
    Department of Bioanalytical Sciences, Genentech, Inc., South San Francisco, California, United States
  • Mauricio Maia
    Department of Bioanalytical Sciences, Genentech, Inc., South San Francisco, California, United States
  • Cynthia Young
    Department of In Vivo Studies, Genentech, Inc., South San Francisco, California, United States
  • Shannon Stainton
    Department of In Vivo Studies, Genentech, Inc., South San Francisco, California, United States
  • Philip E. Hass
    Department of Protein Chemistry, Genentech, Inc., South San Francisco, California, United States
  • Sean B. Joseph
    Department of Preclinical and Translational Pharmacokinetics Genentech, Inc., South San Francisco, California, United States
  • Saileta Prabhu
    Department of Preclinical and Translational Pharmacokinetics Genentech, Inc., South San Francisco, California, United States
  • C. Andrew Boswell
    Department of Preclinical and Translational Pharmacokinetics Genentech, Inc., South San Francisco, California, United States
  • Correspondence: Leslie J. Dickmann, Preclinical and Translational Pharmacokinetics, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080, USA; [email protected]
Investigative Ophthalmology & Visual Science October 2015, Vol.56, 6991-6999. doi:https://doi.org/10.1167/iovs.15-17457
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Leslie J. Dickmann, Victor Yip, Cuiwen Li, Juan Abundes, Mauricio Maia, Cynthia Young, Shannon Stainton, Philip E. Hass, Sean B. Joseph, Saileta Prabhu, C. Andrew Boswell; Evaluation of Fluorophotometry to Assess the Vitreal Pharmacokinetics of Protein Therapeutics. Invest. Ophthalmol. Vis. Sci. 2015;56(11):6991-6999. https://doi.org/10.1167/iovs.15-17457.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose: In this work, we assessed the ability of fluorophotometry to measure the vitreal pharmacokinetics (PK) of fluorescently-labeled ranibizumab in the rabbit after intravitreal injection. We compared these values to those obtained using enzyme-linked immunosorbent assays (ELISA). Data obtained in this study were also compared to historical ranibizumab ocular PK data, either measured in-house or previously published.

Methods: Three individual in vivo studies were performed in New Zealand White rabbits to assess the feasibility of using fluorophotometry to measure rabbit ocular PK of ranibizumab; explore the dynamic range of dosing fluorescently-labeled ranibizumab; and directly compare ranibizumab concentrations and calculated PK parameters measured by vitreal fluorophotometry to those measured using ELISA.

Results: In direct comparisons between fluorophotometry and ELISA, the calculated clearance (CL) values were 0.26 and 0.21 mL/day, the volumes of distribution at steady state (Vss) were 0.80 and 0.94 mL, the half-lives (t1/2) were 3.1 and 2.9 days and the dose normalized areas under the curve (AUC/D) were 4.7 and 3.9 μg·day/mL/μg, respectively. These values fell within the ranges of 0.13 to 0.44 mL/day for CL, 0.5 to 1.8 mL for Vss, 2.8 to 3.5 days for t1/2, and 2.3 to 7.9 μg·day/mL/μg for AUC/D that have been either measured previously in-house or published elsewhere.

Conclusions: Although not suitable for measuring retinal concentrations, fluorophotometry is a valuable, noninvasive method to measure vitreous concentrations of protein therapeutics after intravitreal injection.

Originally developed as a clinical instrument to determine blood–retinal barrier permeability, early cataract formation, tear flow rates, and aqueous humor turnover,13 fluorophotometry (FPT) has become a versatile technique in ocular sciences. Ocular fluorophotometry was developed to further the understanding of ocular pharmacology, physiology, and pathology. Recently, ocular fluorophotometry has been used to assess the ocular pharmacokinetics (PK) of small molecules in preclinical models. In this method, molecules either possess inherent fluorescence or are labeled with a fluorescent dye (usually fluorescein) and administered either systemically or directly into ocular tissue with subsequent measurements of the fluorophore in the transparent ocular tissues. 
Ocular fluorophotometry allows continuous measurement of fluorescence along the central axis of the eye. The dynamic range and lower limits of quantitation of fluorophotometry are similar to other techniques such as mass spectrometry (MS) and ELISA. However, these methods often require invasive and/or terminal sampling. Those options result in the use of a large number of animals and the inability to measure drug concentrations longitudinally in the same animal. Ocular fluorophotometry may offer a noninvasive alternative to measure ocular concentrations and the PK of protein therapeutics and sustained delivery technologies. 
Several groups have successfully used fluorophotometry to measure the ocular PK of nonprotein molecules and to investigate different methods of ocular drug delivery.47 Berezovsky et al.4 used fluorophotometry to evaluate trans-scleral delivery of sodium fluorescein and fluorescein conjugated to 40 and 70 kDa dextrans (FITC-D) in rabbits. In this study, the authors were able to measure retinal/choroidal and vitreous levels of all three molecules after sub-Tenon injection using fluorophotometry and found that the extent of penetration into ocular tissues was dependent on the molecular weight. The concentrations achieved with free fluorescein were much higher in the retina/choroid and vitreous tissues than obtained with 70 kDa FITC-D. Concentrations of the 40 kDa FITC-D were in-between those of sodium fluorescein and the 70 kDa FITC-D. In addition, the ocular concentrations of sodium fluorescein and the 40 and 70 kDa FITC-D in live and euthanized animals were assessed. Ocular tissue concentrations of these molecules were higher in live versus euthanized animals. This result highlights the necessity to conduct ocular PK assessments in vivo. 
Currently, there is no literature describing the use of fluorophotometry in assessing the ocular PK of protein therapeutics such as antibodies, antibody fragments, or fusion proteins. Preliminary results assessing the utility of fluorophotometry to measure ocular PK of rhuFab V2 were presented at a past ARVO conference, but never published (Gaudreault J, et al. IOVS 2002;43:ARVO E-Abstract B798). In addition, the results of PK analysis using fluorophotometry have not been compared with those using MS or ELISA. The objectives of our current study were to address these two gaps by using fluorescently labeled ranibizumab, an anti-VEGF Fab. A comparison of measurement methods was performed to allow for a head-to-head comparison of fluorophotometry with ELISA, the “gold standard” method used for the measurement of protein therapeutics in biological matrices. In addition to this comparison between fluorophotometry and ELISA, the data obtained in this study were also compared to historical ocular PK data obtained with ranibizumab. Our data suggest that the use of fluorophotometry is a valuable, noninvasive method to measure the ocular concentrations of protein therapeutics. 
Materials and Methods
Generation of AlexaFluor 488–Labeled Ranibizumab
AlexaFluor 488 (AF488; Life Technologies, Carlsbad, CA, USA) was chosen for labeling due to its superior brightness, photostability, and pH insensitivity compared to fluorescein. The dye molecule to antibody ratio was kept as close to 1:1 as possible (as assessed by the instructions outlined in Life Technologies manual number MP 10235), although this could vary between different lots of AF488-labeled ranibizumab (AF488-ranibizumab). Ranibizumab (Genentech, South San Francisco, CA, USA) in 0.1M sodium bicarbonate was labeled with a 2-fold molar excess of AF488 NHS ester (product number A-10235; Life Technologies) to create an approximate 1:1 labeling ratio. The excess dye was removed from AF488-ranibizumab using S200 (GE Healthcare, London, UK) size-exclusion chromatography. The eluted peak was pooled, concentrated, dialyzed into PBS, and sterile filtered. Endotoxin levels were assessed using the Pierce LAL chromogenic endotoxin quantitation kit (Life Technologies, Foster City, CA, USA). Endotoxin levels were less than 2 mg/mL. 
In Vivo Dosing and Vitreous Collection
All animal procedures were conducted under a protocol approved by the Institutional Animal Care and Use Committee in an AAALACi-accredited facility in accordance with the Guide for the Care and Use of Laboratory Animals and applicable laws and regulations. New Zealand White rabbits (male and female, >2.5 kg, 3–6 months of age) were used for all in vivo studies. Up to 30 minutes prior to dosing, rabbits were placed in cat bags and topical tropicamide (1%; Akorn Pharmaceuticals, Lake Forest, IL, USA) and/or phenylephrine (2 drops, Akorn Pharmaceuticals) was applied to each eye to dilate the pupil. Rabbits were removed from the cat bags and returned to their cages. For intraocular injections, rabbits were anesthetized with xylazine (3–5 mg/kg, Akorn Pharmaceuticals) approximately 10 minutes prior to ketamine (20–40 mg/kg; Henry Schein, Dublin, OH, USA). Once anesthetized, a topical anesthetic, proparacaine, (Bausch and Lomb, Rochester, NY, USA) was applied to both eyes (2 drops) and buprenorphine (0.02 mg/kg IV bolus, Reckitt Benckiser Healthcare, Berkshire, UK) was given. The dosed eye was opened with a lid speculum and the globe fixated with a Colibri forceps. The eye was then washed with a 1:10 diluted Betadine solution (Henry Schein) and flushed with saline solution (OCuSOFT, Rosenberg, TX, USA). Sterile gauze was used to absorb any pooled solution from the flush. A 28-gauge insulin syringe with a 12-mm needle was used. The needle was inserted approximately 5 mm deep into midvitreous (inferior temporal) and a dose volume not to exceed 50 μL was delivered. Animals that exhibited any regurgitation following injection were not used further in the study, and replaced with another animal that did not display signs of regurgitation. Animals were monitored until they were able to right themselves and then returned to their cages. Rabbits were monitored twice daily while being dosed with buprenorphine. Buprenorphine was administered once prior to recovery from anesthesia, then every 12 hours for 3 days. 
At each scanning time point, the animals were placed under ketamine 20 to 40 mg/kg + xylazine 3 to 5 mg/kg intramuscularly before fluorophotometry scans. For terminal time points, animals were treated with Beuthanasia (Virbac Animal Health, Fort Worth, TX, USA) 4.5 kg/mL intravenously 10 minutes prior to ocular tissue collection. Each eye was dissected, the vitreous (approximately 1.3 mL) was removed with a 1 cm3 syringe with no needle, and placed in an Eppendorf tube and frozen on dry ice until assayed. 
To keep within the dynamic range of the instrument but still deliver the desired amount of protein, AF488-ranibizumab was combined with unlabeled ranibizumab as described in each individual study (Fig. 1). The dosing solution for the first study was prepared by spiking various levels of AF488-ranibizumab (30%, 60%, and 80% of total ranibizumab) with unlabeled ranibizumab to a total amount of 500 μg ranibizumab in PBS per eye. Dosing solutions for subsequent studies were prepared the same way as the initial study but with a 1:1 ratio of AF488-ranibizumab to unlabeled ranibizumab. In the second study, the rabbits were dosed with different amounts of ranibizumab in each group (10, 50, 500 and 1000 μg per eye) with a single 24 hours post dosing termination point. In the third study, rabbits were dosed with the same amount of ranibizumab per eye (500 μg) but terminated at various time points (6, 24, 168, and 336 hours post dosing). 
Figure 1
 
Diagramatic representation of the individual in vivo studies performed herein. (A) In study 1, total ranibizumab concentrations were held constant while the amount of fluorophore varied. Vitreous concentrations were measured via FPT. The right eyes of two animals were dosed for each level of fluorophore examined. (B) In study 2, both ranibizumab and fluorophore were varied while keeping their ratio constant and vitreous concentrations were measured using ELISA and FPT. Two animals were dosed bilaterally for a total of four individual measurements at each dose. (C) In study 3, both ranibizumab and fluorophore were held constant using appropriate concentrations as determined in study 2. Vitreous concentrations and pharmacokinetic (PK) parameters were measured using FPT and ELISA and molecule stability was analyzed via high performance liquid chromatography (HLPC). Two animals were dosed bilaterally for a total of four individual measurements at each time point.
Figure 1
 
Diagramatic representation of the individual in vivo studies performed herein. (A) In study 1, total ranibizumab concentrations were held constant while the amount of fluorophore varied. Vitreous concentrations were measured via FPT. The right eyes of two animals were dosed for each level of fluorophore examined. (B) In study 2, both ranibizumab and fluorophore were varied while keeping their ratio constant and vitreous concentrations were measured using ELISA and FPT. Two animals were dosed bilaterally for a total of four individual measurements at each dose. (C) In study 3, both ranibizumab and fluorophore were held constant using appropriate concentrations as determined in study 2. Vitreous concentrations and pharmacokinetic (PK) parameters were measured using FPT and ELISA and molecule stability was analyzed via high performance liquid chromatography (HLPC). Two animals were dosed bilaterally for a total of four individual measurements at each time point.
Ocular Measurements of AF488-Ranibizumab Using Fluorophotometry
Rabbit eyes were dilated with both topical 1% tropicamide and at least 2 drops of phenylephrine hydrochloride USP 2.5% (with additional phenylephrine if not adequately dilated) 15 to 30 minutes before fluorophotometry measurement. Rabbits were anesthetized with inhalable isoflurane gas at least 5 minutes prior to the measurement at survival time points while at terminal time points, ketamine 20 to 40 mg/kg + xylazine 3 to 5 mg/kg and Beuthanasia 4.5 kg/mL was used as anesthetic. 
After the rabbit had been anesthetized and pupil dilated, ocular measurements of AF488-ranibizumab were done with a Fluorotron Master (OcuMetric, Inc., Mountain View, CA, USA) instrument and software. The animal eye was aligned with the optic head as assessed by the operator viewing window. The fluorescence signal measured was then converted to total ranibizumab concentration based on a standard curve of AF488-ranibizumab test article measured using the cuvette attachment. 
Vitreous Measurements of Ranibizumab and AF488-Ranibizumab Using ELISA
A qualified ligand-binding ELISA was designed to measure the concentrations of AF488-ranibizumab and/or unlabeled ranibizumab in the rabbit vitreous matrix. The assay used recombinant human VEGF-A (Genentech, Inc., South San Francisco, CA, USA) to capture both molecules in rabbit vitreous samples. Bound molecules were detected using an antihuman F(ab')2 conjugated to horseradish peroxidase (HRP, Jackson Immuno Research Laboratories, West Grove, PA, USA), and a peroxidase substrate, 3,3′,5,5′-tetramethylbenzidine, (TMB; Kirkegaard & Perry Laboratories, Gaithersburg, MD, USA) was used for color development. The drug level was quantitated using absorbance spectrophotometry using a microplate reader (Molecular Devices, Sunnyvale, CA, USA). The concentration of total ranibizumab in the study samples was calculated from a standard curve, and the minimum quantifiable concentration (MQC) in rabbit vitreous matrix was 1.5 ng/mL. We found AF488-ranibizumab and unlabeled ranibizumab behaved similarly in this assay (data not shown), and the AF488 did not interfere with binding to either VEGF-A, detection reagent or the measurement of color development using absorbance spectrophotometry. 
Vitreous Chromatography
The vitreous samples collected in study 3 were analyzed by high-performance liquid chromatography (HPLC) with a size-exclusion column (300 × 7.8 mm, BioSep-SEC-s3000; Phenomenex, Torrance, CA, USA) and detected by an in-line fluorescence detector (FLD SEC-HPLC, Agilent Open LAB 1100 HPLC; Agilent Technologies, Santa Clara, CA, USA) with excitation and emission wavelengths of 493 and 517 nm, respectively. Phosphate buffered saline was used as the mobile phase with the isocratic flow rate of 0.5 mL/minutes for 35 minutes. Vitreous samples were diluted with naïve rabbit vitreous to keep within the detection range of the detector, and then further diluted with an equal amount of PBS to reduce the viscosity. The resulting chromatogram profiles were compared to determine the stability of AF488-ranibizumab over the course of the study. The concentration of AF488-ranibizumab was calculated from the standard curve, and the minimum quantifiable concentration in rabbit vitreous matrix was 1.0 ng/mL. The total ranibizumab concentration was calculated from the ratio of AF488 and unlabeled ranibizumab in the study. 
Pharmacokinetic and Statistical Analysis
Vitreous concentration–time profiles were used to estimate PK parameters using noncompartmental analysis (Phoenix WinNonlin, version 6.3; Pharsight Corporation, Sunnyvale, CA, USA). All available data points above the limits of quantitation were used for PK analysis or graphical summary purposes. Clearance was calculated using the equation CL = dose/AUC where dose is known and AUC is measured using the linear trapezoidal method. The volume of distribution at steady state was calculated using the clearance value and the elimination rate constant obtained from the slope of the terminal phase, V = CL/kel. Statistical analysis (2-tailed, unpaired t-test and ANOVA) and graphing was performed using Prism version 6.01 software (GraphPad, La Jolla, CA, USA). 
Results
All intravitreal injections were well tolerated and without complications or in-life findings. For all studies, ranibizumab concentration was calculated based on a standard curve that was generated on the fluorophotometer using either stock solutions or the individual dosing solutions. In previous studies (Supplementary Fig. S1), similar results were obtained with a standard curve generated in either PBS or rabbit vitreous humor. Therefore, standard curves used in studies described in this manuscript were generated in PBS. 
Initial Feasibility Study With AF488-Ranibizumab (Study 1)
To determine the feasibility of using fluorophotometry to assess the PK of AF488-ranibizumab administered intravitreally, an initial study was performed as illustrated in Figure 1A. The total amount of administered ranibizumab remained constant between treatment groups while the level of AF488 dye varied. This allowed initial investigation of the dynamic range of the instrument while keeping the total protein load constant. An example of the raw fluorophotometry depth scans over the course of 14 days is shown in Figure 2A. The broad vitreous peak between approximately 0 to 20 mm was observed over the entire time period and for all AF488 levels (Fig. 2A, Supplementary Fig. S2). Two methods were employed to translate the raw fluorescence signal to the concentration of ranibizumab. The first was the peak method that utilized the maximum peak measured between 0 and 20 mm (Figs. 2B, 3). The second method utilized the total area under the curve (AUC) between 0 and 20 mm (Fig. 2B, Supplementary Fig. S3). As shown in Figure 3A and Supplementary Figure S4A, the fluorescence signal measured using the fluorophotometer showed a first order loss over the course of 14 days for all fluorescence levels evaluated. Signal normalized to fluorescence and to an AF488-ranibizumab standard curve (Figs. 3B, 3C, Supplementary Figs. S4B, S4C) also displayed first-order kinetics. Normalized ranibizumab concentrations were used to calculate the PK parameters (Table 1) for individual animals for both the peak and AUC methods. Animal-to-animal variability was low for all PK parameters as evidenced by the standard deviations in Table 1. Using the peak method, the calculated vitreous clearance (CL) of ranibizumab was 0.25 ± 0.02 mL/day, the volume of distribution at steady state (Vss) was 1.1 ± 0.13 mL, the half-life (t1/2) was 3.2 ± 0.25 days, and the AUC was 1944 ± 114 μg·day/mL. Using the AUC method, the calculated vitreous CL of ranibizumab was 0.33 ± 0.06 mL/day, the Vss was 1.5 ± 0.32 mL, the t1/2 was 3.3 ± 0.12 days, and the AUC was 1540 ± 372 μg·day/mL. As shown in Table 1, differences in CL, Vss, and AUC values were statistically significant (P value < 0.05) between the peak method and AUC method, whereas t1/2 values were not. However, all values calculated were within the ranges observed historically, suggesting that both the peak and AUC methods are viable approaches for normalizing the fluorescence signal obtained with fluorophotometry. For subsequent analysis, only the peak method was used with fluorophotometry. 
Figure 2
 
Depth scans of fluorophotometric signals. (A) Representative depth scan of raw signals from animal 2B over the course of 14 days. The preread measurement is not visible due to the low level of fluorescence of the preread scan. (B) Fluorophotometry depth scan exemplifying the peak (dashed circle) and AUC (solid outline) methods. The area under the curve method utilized total area under the curve from 0 to 20 mm, while the peak method used the maximum signal between 0 and 20 mm.
Figure 2
 
Depth scans of fluorophotometric signals. (A) Representative depth scan of raw signals from animal 2B over the course of 14 days. The preread measurement is not visible due to the low level of fluorescence of the preread scan. (B) Fluorophotometry depth scan exemplifying the peak (dashed circle) and AUC (solid outline) methods. The area under the curve method utilized total area under the curve from 0 to 20 mm, while the peak method used the maximum signal between 0 and 20 mm.
Figure 3
 
Concentration time profiles for the peak method of analysis. Figures 1A through 3B denote individual animals where only the right eye was dosed. (A) Fluorophotometric signal versus time based on the maximum signal between 0 and 20 mm. (B) Signal normalized to percentage of conjugated AF488. (C) Concentrations calculated from the AF488-ranibizumab standard curve.
Figure 3
 
Concentration time profiles for the peak method of analysis. Figures 1A through 3B denote individual animals where only the right eye was dosed. (A) Fluorophotometric signal versus time based on the maximum signal between 0 and 20 mm. (B) Signal normalized to percentage of conjugated AF488. (C) Concentrations calculated from the AF488-ranibizumab standard curve.
Table 1
 
Pharmacokinetic Parameters Obtained by Measuring AF488-Labeled Ranibizumab Using FPT
Table 1
 
Pharmacokinetic Parameters Obtained by Measuring AF488-Labeled Ranibizumab Using FPT
Exploring the Dynamic Range and Sensitivity (Study 2)
To further explore the dynamic range and sensitivity of fluorophotometry, a second study was performed as outlined in Figure 1B. In this study, animals received 0.01 to 1.0 mg/eye ranibizumab of which 50% was fluorescently labeled. Therefore, both total protein and total fluorescence varied between treatment groups, although the ratio remained the same. After taking a fluorophotometric reading at 24 hours, animals were euthanized and the vitreous humor was collected for measurement of the concentrations of ranibizumab with ELISA. From previous experiments (Fig. 2 and data not shown), the 24-hour time point was found to be a long enough time period after dosing to achieve near homogeneous distribution of a Fab molecule throughout the visual axis. As shown in Figure 4 and Table 2, vitreous concentrations of ranibizumab measured via fluorophotometry and ELISA were comparable at the 10 and 500 μg/eye doses. Concentrations of ranibizumab were 7.14 ± 1.04 and 5.69 ± 0.80 μg/mL for the 10 μg/eye dose and 364 ± 46 and 367 ± 14 μg/mL for the 500 μg/eye dose for fluorophotometry and ELISA, respectively. For the 50 μg/eye dose, concentrations of ranibizumab were statistically significantly different between fluorophotometry and ELISA (P < 0.05) with the concentrations of 48.0 ± 4.4 and 37.4 ± 4.0 μg/mL, respectively. At the 1000 μg/eye dose, the concentrations of ranibizumab determined using ELISA and fluorophotometry diverged significantly. Vitreous concentrations were 532 ± 63 and 694 ± 89 μg/mL for fluorophotometry and ELISA measurements, respectively, and this difference was statistically significant (P < 0.05). 
Figure 4
 
Direct comparison of FPT and ELISA 24 hours after dosing (study 2). After fluorophotometric readings, vitreous humor was immediately collected for total ranibizumab measurement using ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). Each asterisk indicates a statistically significant difference between the two groups at that time point. *P < 0.05 using a 2-tailed, unpaired t-test.
Figure 4
 
Direct comparison of FPT and ELISA 24 hours after dosing (study 2). After fluorophotometric readings, vitreous humor was immediately collected for total ranibizumab measurement using ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). Each asterisk indicates a statistically significant difference between the two groups at that time point. *P < 0.05 using a 2-tailed, unpaired t-test.
Table 2
 
Concentrations of Total Ranibizumab Measured in Rabbit Vitreous Humor by FPT or ELISA 24 Hours After Dosing
Table 2
 
Concentrations of Total Ranibizumab Measured in Rabbit Vitreous Humor by FPT or ELISA 24 Hours After Dosing
Direct Pharmacokinetic Comparison Between Fluorophotometry and ELISA and Comparison to Historical Data (Study 3)
For a direct comparison of PK parameters determined using fluorophotometry and ELISA, a third study was performed as outlined in Figure 1CA 500 μg/eye dose of total ranibizumab (1:1 AF488-ranibizumab to unlabeled ranibizumab) was given intravitreally. Then, ranibizumab concentrations that were measured by fluorophotometry and by ELISA were used to generate PK parameters using noncompartmental analysis. The results of this study are shown in Figure 5 and Table 3. Vitreous CL values for ranibizumab were 0.26 and 0.21 mL/day, Vss values were 0.80 and 0.94 mL, and t1/2 values were 3.1 and 2.9 days for concentrations measured using fluorophotometry and ELISA, respectively. The values of AUC were normalized by dose (AUC/D) for comparison to historical data and were 4.7 and 3.9 day·μg/mL/μg for fluorophotometry and ELISA, respectively. To better assess whether the PK values obtained from the studies herein were within the expected range, they were compared to historical values that were either obtained at Genentech over the course of several years or identified from the literature. Nine total historical studies were used in this evaluation. Seven PK studies in rabbits were performed at Genentech using ranibizumab, one of which had been previously published.8 In the nine historical studies, the mean and standard deviation values for CL, Vss, t1/2, and AUC/D of ranibizumab were 0.30 ± 0.09 mL/day, 1.3 ± 0.4 mL, 3.2 ± 0.3 days, and 3.8 ± 1.9 day·μg/mL/μg, respectively. The range of values for CL, Vss, t1/2, and AUC/D of ranibizumab were 0.13 to 0.44 mL/day, 0.5 to 1.8 mL, 2.8 to 3.5 days, and 2.3 to 7.9 day·μg/mL/μg, respectively. The PK parameters calculated in studies 1 and 3 herein using concentrations of ranibizumab determined by fluorophotometry or ELISA were within the ranges measured in previous studies. 
Figure 5
 
Direct comparison of FPT and ELISA measurements from study 3. Concentration versus time profiles over the course of 14 days comparing FPT and ELISA data. After FPT readings were obtained, vitreous humor was immediately collected for total ranibizumab measurement using an ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). There was no statistically significant difference between the ELISA and FPT measurements at each individual time point as assessed using an unpaired t-test.
Figure 5
 
Direct comparison of FPT and ELISA measurements from study 3. Concentration versus time profiles over the course of 14 days comparing FPT and ELISA data. After FPT readings were obtained, vitreous humor was immediately collected for total ranibizumab measurement using an ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). There was no statistically significant difference between the ELISA and FPT measurements at each individual time point as assessed using an unpaired t-test.
Table 3
 
Comparison of Ranibizumab PK Parameters Measured in the Current Study and Those Measured in Prior Studies
Table 3
 
Comparison of Ranibizumab PK Parameters Measured in the Current Study and Those Measured in Prior Studies
Assessment of AF488-Ranibizumab Stability
The stability of the AF488-ranibizumab was assessed from rabbit vitreous samples obtained from study 3. Exposure to the ocular environment and ambient light over time could potentially lead to instability of the Fab-dye conjugate or decreased dye sensitivity. As shown in Figure 6, AF488-ranibizumab analyzed from vitreous samples 7 or 14 days after dosing showed similar peak retention times compared to the dosing solution with no indication of free dye or unexpected protein-dye fragments. This result suggested that the Fab-dye conjugation remained intact and was stable in the vitreous. When normalized to total ranibizumab concentration using a standard curve generated by HPLC-fluorescence, concentrations were comparable to those measured using fluorophotometry and ELISA. This outcome suggested that dye sensitivity was not affected (Supplementary Fig. S3). 
Figure 6
 
Stability of AF488-labeled ranibizumab. Ranibizumab-AF488 was measured via high performance liquid chromatography fluorescence as described in the Materials and Methods section. The initial dosing solution was diluted 8-fold in PBS for this comparison to be within the dynamic range of the detector and the expected concentration of the vitreous samples. Vitreous samples from days 7 and 14 were each diluted 1:1 with PBS for this analysis. The chromatogram is representative of all animals in study 3.
Figure 6
 
Stability of AF488-labeled ranibizumab. Ranibizumab-AF488 was measured via high performance liquid chromatography fluorescence as described in the Materials and Methods section. The initial dosing solution was diluted 8-fold in PBS for this comparison to be within the dynamic range of the detector and the expected concentration of the vitreous samples. Vitreous samples from days 7 and 14 were each diluted 1:1 with PBS for this analysis. The chromatogram is representative of all animals in study 3.
Discussion
The measurement of ocular drug concentrations and assessment of ocular PK in preclinical species is an important aspect of ophthalmologic drug development. It allows for the prediction of human ocular PK and thus helps guide clinical dosing strategies. Measuring ocular drug concentrations generally relies on terminal euthanasia of the animal in order to obtain ocular matrices. Although an aqueous tap is a useful technique that can be done without killing the animal, the invasive nature of this method precludes it from being performed more than every other week in the rabbit. Finding an alternative to highly invasive ocular sampling and collection techniques would not only advance the fundamental practice of 3Rs (reduction, refinement, and replacement), it would also improve scientific practice as longitudinal measurements could be obtained from the same animal. 
Ocular fluorophotometry has the unique ability to excite a fluorophore in the eye and to measure the emission from this fluorophore. Fluorescence is only quantified along the central visual axis of the eye. Thus, it is limited to one-dimensional scans. However, fluorophotometry offers a fast, safe, and sensitive method of measurement in the eyes of live animals. Although other technologies exist, such as MRI,9,10 that can provide a three-dimensional image of eyes in vivo, they tend to be much more expensive, and have limited sensitivity and resolution with extended imaging times. Additionally, the physicochemical properties of contrast agents currently available for MRI would be expected to affect the PK of protein therapeutics, and would therefore be inappropriate for application in this context. 
There are caveats and assumptions that need to be addressed when using ocular fluorophotometry for assessment of PK. One assumption is that molecules will distribute homogenously through the vitreous and central visual axis such that the measurements taken via fluorophotometry are representative of concentrations throughout the entire vitreous. This assumption may not be valid for molecules or long-acting delivery technologies that may form a depot outside of the central visual axis, or for which distribution is markedly slower. Thus, the concentration measured using fluorophotometry may not be indicative of the entire vitreous. In this case, other techniques may offer a more comprehensive view of the molecular distribution and localized concentration throughout the entire eye. 
It may also be difficult, if not impossible, to accurately measure retinal concentrations after an intravitreal dose of a fluorescently labeled protein therapeutic. The reason for this is inherent in the fluorophotometer itself. As the detection region of fluorescence measurement (approximately 1 mm)3 is much larger than the thickness of the rabbit retina (0.2 mm),11 there will likely always be overlap in the readings between retina and vitreous. This overlap may not be of great concern for molecules administered via sub-Tenon injections where fluorescence in the vitreous is quite low.4,7 However, it will be of concern for molecules administered intravitreally where vitreous fluorescence levels could be orders of magnitude higher than retinal levels and thus “bleed” into retinal measurements. This tailing effect, or the tendency of highly fluorescent structures to influence the fluorescence of areas immediately adjacent to them, has been previously observed.3,12 Because of this phenomenon, measurement of ranibizumab in the retina using fluorophotometry was not attempted in this study. 
In these studies, drug concentrations were not assessed in the aqueous humor. In theory, the same experimental techniques employed here for the vitreous humor can also be used to assess the concordance between fluorophotometry and ELISA measurements in the aqueous humor. As seen in Figure 2B, there is a fluorescence signal in the depth scan area that corresponds to the anterior chamber (approximately 22–28 mm). Parameters of PK using fluorophotometry were not assessed for aqueous humor in this study, due to low fluorescence signal in the aqueous humor by day 7. Fluorescence concentrations would thus have to be optimized in order to measure aqueous humor PK. 
Caveats aside, this study demonstrated the utility of using fluorophotometry to assess vitreous PK of protein therapeutics. When using fluorophotometry, it is important to keep the fluorescence signal within the dynamic range of the instrument over the appropriate time course. The starting fluorescence level will likely be dependent on the dye molecule used (e.g., AF488, fluorescein, Oregon green), the dye to protein molecule ratio, and the protein therapeutic molecule used in the study. Although several fluorescein-like fluorophores can be detected by this particular instrument, each fluorophore has specific excitation and emission spectra, pH sensitivities, and extinction coefficients. Prestudy standard curve analysis was and should be used to guide the precise amount of labeled versus unlabeled protein used in the fluorophotometric experiments. In our current study, the dye-to-Fab ratio was kept as close to one to one as possible as signal quenching has been observed for molecules with high dye to protein ratios (data not shown). However, our data (studies 1 and 2 herein) do indicate flexibility in the extent of dye labeling and the amount of labeled molecule used. Our results from studies 2 and 3 indicate that there is good agreement between the concentrations measured by fluorophotometric and ELISA techniques when the level of fluorescence is kept within the dynamic range of the instrument. In study 2, the main divergence between fluorophotometric and ELISA values appeared at fluorescence concentrations that were above the dynamic range of the instrument for this particular fluorophore–protein combination. 
Based on Table 3, it appears that the PK parameters calculated using the peak method and AUC method gave comparable results. Because the peak method showed less variability in calculated CL, Vss, and AUC values, it was used for subsequent analyses. The reason for the increased variability using the AUC method is currently not fully understood, and the limited data comparing the two methods make any theory difficult to support. However, as seen in Figure 2 and Supplementary Figure S2, the fluorometric peak shapes tend to be quite variable at the early time points (up to 48 hours). This may be due to the exact location of administration and subsequent diffusion of the molecule through the vitreous, parameters that are likely to be slightly different from eye to eye. 
Ranibizumab PK parameter estimates from studies 1 and 3 are also in good agreement with those that were measured previously at Genentech or were published. Clearance, Vss, t1/2, and AUC/D values fell within the range of those measured in previous studies. Although published PK parameters for ranibizumab measured by groups outside of Genentech are limited, the t1/2 values reported in these studies are in line with those measured in the current studies using both ELISA and fluorophotometry. One limitation of the current investigation is the lack of study-to-study variability using fluorophotometry with direct comparison to ELISA. While the average of several studies is likely a better comparison to use against the historical data, the authors felt that it did not justify the use of more animals. We do believe, however, that nonterminal PK studies using fluorophotometry in higher order species such as mini-pig or cynomolgus monkey are worth investigating as these species are generally considered to be better ocular models for humans. In summary, the present study indicates that fluorophotometry can be a useful tool for assessing the PK of protein therapeutics in the rabbit vitreous. Although the initial cost of instrument purchase (approximately $65,000) could be a limiting factor for some labs, the cost savings in both animal usage and work time could likely offset this. The need for time-consuming assay development, especially at the drug discovery stage, would be completely negated. Fluorophotometry also allows for longitudinal measurements in live animals and can reduce animal usage by up to 70% for standard ocular PK studies. 
Acknowledgments
The authors thank Harbeen Grewal and Mausumi Debnath from Anshin BioSolutions for editorial assistance. 
Disclosure: L.J. Dickmann, Genentech, Inc. (I, E); V. Yip, Genentech, Inc. (I, E); C. Li, Genentech, Inc. (I, E); J. Abundes Jr, Genentech, Inc. (I, E); M. Maia, Genentech, Inc. (I, E); C. Young, Genentech, Inc. (I, E); S. Stainton, Genentech, Inc. (I, E); P.E. Hass, Genentech, Inc. (I, E); S.B. Joseph, Genentech, Inc. (I, E); S. Prabhu, Genentech, Inc. (I, E); C.A. Boswell, Genentech, Inc. (I, E) 
References
Knudsen LL. Ocular fluorophotometry in human subjects and in swine - with particular reference to long-term pharmacokinetics. Acta Ophthalmol Scand Suppl. 2002; 80: 6–24.
Spalton DJ. Ocular fluorophotometry. Br J Ophthalmol. 1990; 74: 431–432.
McCarey B. Fluorophotometry for pharmacokinetic assessment. In: Edelhauser UKaH, ed. Drug Product Development for the Back of the Eye. New York: Springer; 2011: 47–84.
Berezovsky DE, Patel SR, McCarey BE, Edelhauser HF. In vivo ocular fluorophotometry: delivery of fluoresceinated dextrans via transscleral diffusion in rabbits. Invest Ophthalmol Vis Sci. 2011; 52: 7038–7045.
Ghate D, Brooks W, McCarey BE, Edelhauser HF. Pharmacokinetics of intraocular drug delivery by periocular injections using ocular fluorophotometry. Invest Ophthalmol Vis Sci. 2007; 48: 2230–2237.
Kizhakkethara I, Li X, el-Sayed S, et al. Noninvasive monitoring of intraocular pharmacokinetics of daunorubicin using fluorophotometry. Int Ophthalmol. 1995; 19: 363–367.
Lee SJ, Kim ES, Geroski DH, McCarey BE, Edelhauser HF. Pharmacokinetics of intraocular drug delivery of Oregon green 488-labeled triamcinolone by subtenon injection using ocular fluorophotometry in rabbit eyes. Invest Ophthalmol Vis Sci. 2008; 49: 4506–4514.
Gaudreault J, Fei D, Beyer JC, et al. Pharmacokinetics and retinal distribution of ranibizumab, a humanized antibody fragment directed against VEGF-A, following intravitreal administration in rabbits. Retina. 2007 ; 27: 1260–1266.
Kim H, Robinson MR, Lizak MJ, et al. Controlled drug release from an ocular implant: an evaluation using dynamic three-dimensional magnetic resonance imaging. Invest Ophthalmol Vis Sci. 2004; 45: 2722–2731.
Li SK, Lizak MJ, Jeong EK. MRI in ocular drug delivery. NMR Biomed. 2008; 21: 941–956.
Muraoka Y, Ikeda HO, Nakano N, et al. Real-time imaging of rabbit retina with retinal degeneration by using spectral-domain optical coherence tomography. PLoS One. 2012; 7: e36135.
Zeimer RC, Blair NP, Cunha-Vaz JG. Pharmacokinetic interpretation of kinetic vitreous fluorophotometry. Graefes Arch Clin Exp Ophthalmol. 1985; 222: 222.
Ahn SJ, Ahn J, Park S, et al. Intraocular pharmacokinetics of ranibizumab in vitrectomized versus nonvitrectomized eyes. Invest Ophthalmol Vis Sci. 2014; 55: 567–573.
Bakri SJ, Snyder MR, Reid JM, Pulido JS, Ezzat MK, Singh RJ. Pharmacokinetics of intravitreal ranibizumab (Lucentis). Ophthalmology. 2007 ; 114: 2179–2182.
Figure 1
 
Diagramatic representation of the individual in vivo studies performed herein. (A) In study 1, total ranibizumab concentrations were held constant while the amount of fluorophore varied. Vitreous concentrations were measured via FPT. The right eyes of two animals were dosed for each level of fluorophore examined. (B) In study 2, both ranibizumab and fluorophore were varied while keeping their ratio constant and vitreous concentrations were measured using ELISA and FPT. Two animals were dosed bilaterally for a total of four individual measurements at each dose. (C) In study 3, both ranibizumab and fluorophore were held constant using appropriate concentrations as determined in study 2. Vitreous concentrations and pharmacokinetic (PK) parameters were measured using FPT and ELISA and molecule stability was analyzed via high performance liquid chromatography (HLPC). Two animals were dosed bilaterally for a total of four individual measurements at each time point.
Figure 1
 
Diagramatic representation of the individual in vivo studies performed herein. (A) In study 1, total ranibizumab concentrations were held constant while the amount of fluorophore varied. Vitreous concentrations were measured via FPT. The right eyes of two animals were dosed for each level of fluorophore examined. (B) In study 2, both ranibizumab and fluorophore were varied while keeping their ratio constant and vitreous concentrations were measured using ELISA and FPT. Two animals were dosed bilaterally for a total of four individual measurements at each dose. (C) In study 3, both ranibizumab and fluorophore were held constant using appropriate concentrations as determined in study 2. Vitreous concentrations and pharmacokinetic (PK) parameters were measured using FPT and ELISA and molecule stability was analyzed via high performance liquid chromatography (HLPC). Two animals were dosed bilaterally for a total of four individual measurements at each time point.
Figure 2
 
Depth scans of fluorophotometric signals. (A) Representative depth scan of raw signals from animal 2B over the course of 14 days. The preread measurement is not visible due to the low level of fluorescence of the preread scan. (B) Fluorophotometry depth scan exemplifying the peak (dashed circle) and AUC (solid outline) methods. The area under the curve method utilized total area under the curve from 0 to 20 mm, while the peak method used the maximum signal between 0 and 20 mm.
Figure 2
 
Depth scans of fluorophotometric signals. (A) Representative depth scan of raw signals from animal 2B over the course of 14 days. The preread measurement is not visible due to the low level of fluorescence of the preread scan. (B) Fluorophotometry depth scan exemplifying the peak (dashed circle) and AUC (solid outline) methods. The area under the curve method utilized total area under the curve from 0 to 20 mm, while the peak method used the maximum signal between 0 and 20 mm.
Figure 3
 
Concentration time profiles for the peak method of analysis. Figures 1A through 3B denote individual animals where only the right eye was dosed. (A) Fluorophotometric signal versus time based on the maximum signal between 0 and 20 mm. (B) Signal normalized to percentage of conjugated AF488. (C) Concentrations calculated from the AF488-ranibizumab standard curve.
Figure 3
 
Concentration time profiles for the peak method of analysis. Figures 1A through 3B denote individual animals where only the right eye was dosed. (A) Fluorophotometric signal versus time based on the maximum signal between 0 and 20 mm. (B) Signal normalized to percentage of conjugated AF488. (C) Concentrations calculated from the AF488-ranibizumab standard curve.
Figure 4
 
Direct comparison of FPT and ELISA 24 hours after dosing (study 2). After fluorophotometric readings, vitreous humor was immediately collected for total ranibizumab measurement using ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). Each asterisk indicates a statistically significant difference between the two groups at that time point. *P < 0.05 using a 2-tailed, unpaired t-test.
Figure 4
 
Direct comparison of FPT and ELISA 24 hours after dosing (study 2). After fluorophotometric readings, vitreous humor was immediately collected for total ranibizumab measurement using ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). Each asterisk indicates a statistically significant difference between the two groups at that time point. *P < 0.05 using a 2-tailed, unpaired t-test.
Figure 5
 
Direct comparison of FPT and ELISA measurements from study 3. Concentration versus time profiles over the course of 14 days comparing FPT and ELISA data. After FPT readings were obtained, vitreous humor was immediately collected for total ranibizumab measurement using an ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). There was no statistically significant difference between the ELISA and FPT measurements at each individual time point as assessed using an unpaired t-test.
Figure 5
 
Direct comparison of FPT and ELISA measurements from study 3. Concentration versus time profiles over the course of 14 days comparing FPT and ELISA data. After FPT readings were obtained, vitreous humor was immediately collected for total ranibizumab measurement using an ELISA as outlined in the Materials and Methods section. Time points represent average and standard deviation of n = 4 eyes (two animals dosed bilaterally). There was no statistically significant difference between the ELISA and FPT measurements at each individual time point as assessed using an unpaired t-test.
Figure 6
 
Stability of AF488-labeled ranibizumab. Ranibizumab-AF488 was measured via high performance liquid chromatography fluorescence as described in the Materials and Methods section. The initial dosing solution was diluted 8-fold in PBS for this comparison to be within the dynamic range of the detector and the expected concentration of the vitreous samples. Vitreous samples from days 7 and 14 were each diluted 1:1 with PBS for this analysis. The chromatogram is representative of all animals in study 3.
Figure 6
 
Stability of AF488-labeled ranibizumab. Ranibizumab-AF488 was measured via high performance liquid chromatography fluorescence as described in the Materials and Methods section. The initial dosing solution was diluted 8-fold in PBS for this comparison to be within the dynamic range of the detector and the expected concentration of the vitreous samples. Vitreous samples from days 7 and 14 were each diluted 1:1 with PBS for this analysis. The chromatogram is representative of all animals in study 3.
Table 1
 
Pharmacokinetic Parameters Obtained by Measuring AF488-Labeled Ranibizumab Using FPT
Table 1
 
Pharmacokinetic Parameters Obtained by Measuring AF488-Labeled Ranibizumab Using FPT
Table 2
 
Concentrations of Total Ranibizumab Measured in Rabbit Vitreous Humor by FPT or ELISA 24 Hours After Dosing
Table 2
 
Concentrations of Total Ranibizumab Measured in Rabbit Vitreous Humor by FPT or ELISA 24 Hours After Dosing
Table 3
 
Comparison of Ranibizumab PK Parameters Measured in the Current Study and Those Measured in Prior Studies
Table 3
 
Comparison of Ranibizumab PK Parameters Measured in the Current Study and Those Measured in Prior Studies
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×