Preclinical PET Study of Intravitreal Injections

Abstract

Purpose: This work aimed at describing the time course of vitreous clearance through the use of positron emission tomography (PET) as a noninvasive tool for pharmacokinetic studies of intravitreal injection.

Methods: The pharmacokinetic profile of intravitreal injections of molecules labeled with ¹⁸Fluorine (¹⁸F) was evaluated in adult Sprague Dawley rats by using a dedicated small-animal PET/computed tomography scanner. Different conditions were studied: three molecules radiolabeled with ¹⁸F (¹⁸F-FDG, ¹⁸F-NaF, and ¹⁸F-Choline), three volumes of intravitreal injections (7, 4, and 2 μL), and absence or presence of eye inflammation (uveitis).

Results: Our results showed that there are significant pharmacokinetic differences among the radiolabeled molecules studied but not among the injected volumes. The presence or absence of uveitis was an important factor in vitreous clearance, since the elimination of the drug was clearly increased when this condition is present.

Conclusions: Intravitreal pharmacokinetic studies based on the use of dedicated PET imaging can be of potential interest as noninvasive tools in ophthalmic drug development in small animals.

To date, most topical and systemic drugs have not achieved adequate therapeutic levels in the vitreous, mainly owing to the existence of different physiological barriers.¹ On one hand, topically instilled drugs are diluted by the tear film, thus causing significant drug loss in the lachrymal flow,² and furthermore their physicochemical characteristics must be adequate to cross the cornea.³ On the other hand, the blood–retinal barrier (BRB), which comprises the retinal pigment epithelium and the tightly sealed walls of the retinal capillaries, complicates the arrival of systemic drugs to the vitreous.⁴ For these reasons, intravitreal administration has become an effective way to deliver drugs to the vitreous cavity, allowing high drug concentrations.⁵ 

To achieve a sustained therapeutic drug concentration in the vitreous, the frequency of administration should be based on the half-life of the drug (t_1/2). Regarding this question, several in vitro models have been proposed for the study of intravitreal pharmacokinetics, which take into account all aspects of the ocular anatomy and physiology.^6–9 However, one aspect that should be taken into consideration in the in vitro pharmacokinetic studies is the absence of convection,^10–13 even though the principal mechanism of transport through the vitreous is diffusion, and convection does not play a relevant role in the kinetics of small molecules. Other issues such as protein binding, melanin binding, drug metabolism, or active transport are usually not taken into account in the in vitro studies.^8,14 On the other hand, in vivo classical pharmacokinetic studies of intravitreal injections are limited because invasive techniques are involved.^15,16 

In recent years, molecular imaging techniques have become a turning point for the development and pharmacokinetic study of new drugs. These techniques involve noninvasive procedures in order to significantly decrease the number of animals used by increasing the number of measurements on each animal.^17,18 In particular for the field of intravitreal drugs, single photon emission computed tomography and magnetic resonance image (MRI) have been the most commonly used imaging techniques, mainly to study pharmacokinetics^14,19 and the release of drugs from implants and liposomes.^20–22 

However, in pharmacokinetic studies performed with MRI, the molecules used for the labeling of the drug usually have very high molecular weight, which can alter the properties of the original drug.²¹ 

The use of positron emission tomography (PET) has made it possible to label drugs with small β-emitting radioisotopes.²³ Current integrated PET/computed tomography (CT) scanners allow visualization of radiolabeled molecules by using a direct and noninvasive methodology, and the follow-up of the same subject over time to determine the pharmacokinetic properties of intravitreal injections.^24–26 

Different radionuclides can be used to elaborate radiotracers for PET scanning. The most commonly used radionuclides are typically isotopes with short half-lives such as ¹¹C, ¹³N, ¹⁵O, ¹⁸F, ⁶⁸Ga, ⁸²Rb, or with longer half-lives such as ¹²⁴I or ⁸⁹Zr. ¹⁸F is one of the most widely used because it is easily produced with a cyclotron, its positron energy of emission is 0.64 MeV, it is safe for patients, and it allows to obtain images with high resolution. Moreover, its half-life is long enough to be able to produce commercially manufactured fluorinated radiotracers at off-site locations and to be shipped to imaging services. In practice, ¹⁸F radionuclide is linked to different molecules to achieve selective transport and distribution.²⁷ 

Drug clearance in the vitreous can be influenced by various factors that include molecular weight, physicochemical properties of the drug, surgical procedure, injected volumes, and presence of ocular inflammation.¹ Also, the mechanisms of membrane transport and plasmatic clearance can highly influence the distribution and elimination of drugs after intravitreal administration. For this reason, fluorodeoxyglucose (¹⁸F-FDG), ¹⁸F-choline (¹⁸F-Choline), and ¹⁸F–sodium fluoride (¹⁸F-NaF) were selected in our study because of their different molecular weight, polarity, and transport mechanism across biological membranes. The aim of the present work was to study the effect of some of these factors on the vitreous clearance by using dedicated PET/CT imaging techniques for in vivo studies in rats. 

Materials and Methods

Our work was designed as an experimental study in rats scanned in a dedicated PET/CT system after intravitreal injections of different radiolabeled molecules, different volumes, and absence/presence of inflammatory eye disease (uveitis). 

Animals

This study was carried out on male adult Sprague Dawley rats with an average weight of 300 g, supplied by the animal facility of the University of Santiago de Compostela (Santiago de Compostela, Spain). During the experiments, the animals were kept in individual cages with free access to food and water in a room under controlled temperature (22°C ± 1°C) and humidity (60% ± 5%) and with day–night cycles regulated by artificial light (12/12 hours). The animals were treated as indicated in the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and according to the guidelines for laboratory animals.^28,29 Experiments were approved by the Galician Network Committee for Ethical Research and followed the Spanish and European Union (EU) rules (86/609/CEE, 2003/65/CE, 2010/63/EU, RD 1201/2005, and RD53/2013). 

Intravitreal Injection Procedure

Intravitreal injection was performed according to the procedure described previously by Chiu et al.³⁰ Firstly, the animals were placed in a gas chamber containing 2% isoflurane in oxygen. When unconscious, the animals were removed from the chamber but kept under anesthesia with a mask (1.5% isoflurane in oxygen). The procedure was initiated by applying one drop of topical anesthesia (Colircusi Anestesico Doble: tetracaine 1 mg/mL and oxybuprocaine 4 mg/mL) on the eye followed by mydriatic eye drops (phenylephrine 100 mg/mL [Colircusi Fenilefrina] and tropicamide 10 mg/mL [Colircusi Tropicamide]) to visualize the eye fundus. Thereafter, radiolabeled molecules were injected into the vitreous through the pars plana by using a Hamilton syringe with a 34-G needle. The injection procedure was performed with a surgical microscope (Takagi OM-5 220-2; Takagi, Tokyo, Japan). Pictures of the procedure were taken by means of a digital camera (Nikon D-200; Nikon, Tokyo, Japan) attached to the microscope. Eyes with lens damage, or with significant bleeding when the intravitreal injection was made, were discarded from the study. 

Experiments

The experiments were carried out by using intravitreal injections with three radiolabeled molecules and three different injection volumes, in healthy eyes and in eyes with lipopolysaccharide (LPS)–induced uveitis. 

Effect of the Type of Injected Radiolabeled Molecules

Three different molecules were labeled with ¹⁸F to evaluate the intravitreal pharmacokinetics. The radiolabeled molecules to be injected were ¹⁸F-NaF, ¹⁸F-FDG, and ¹⁸F-Choline, with molecular weights of 41, 182, and 122 g/mol, respectively (Fig. 1). 

Figure 1

 

Chemical structure of (A) ¹⁸F-Choline, (B) ¹⁸F-FDG, and (C) ¹⁸F-NaF.

The radioisotope ¹⁸F⁻ was obtained from the nuclear reaction ¹⁸O (proton, neutron) carried out in our PET Trace 800 cyclotron, according to the method described by Saha.³¹ The radiosynthesis of ¹⁸F-Na was made with a carbonate-type anion-exchange resin column, in such a way that the ¹⁸F⁻ is retained into the column and it is recovered as ¹⁸F–sodium fluoride by elution with potassium carbonate solution. ¹⁸F-FDG and ¹⁸F-Choline were produced on a TRACERlab MX synthesizer (GE Healthcare, Waukesah, WI, USA) by using cassettes and reagent kits from ABX (Advanced Biochemical Compounds, Radeberg, Germany). The nucleophilic substitution standard method was used in the case of ¹⁸F-FDG and for the reaction of 18F-fluoromethyl triflate with dimethylethanolamine on a Sep-Pak column used in the case of ¹⁸F-Choline.^32,33 

All procedures to obtain radiolabeled molecules were performed under good-manufacturing-practice conditions following the specific standards of European Pharmacopoeia.³⁴ The purity and stability quality control requirements were undertaken via high-pressure liquid chromatography/ion chromatography (930 Compact IC Flex con; Metrohm AG, Herisau, Switzerland) and thin layer chromatography. Osmolality (mOsm/kg) and pH were determined with a vapor pressure osmometer (VAPRO 5520; ELITECH Group, Paris, France) and a pH meter (WTW inoLab; WTW, Weilheim, Germany). 

Effect of the Injected Volumes

The effect of the injected volume on the intravitreal pharmacokinetics of the abovementioned molecules was evaluated by using three different volumes: 2, 4, and 7 μL. 

Effect of the Presence of Inflammation

Intravitreal pharmacokinetics was assessed in a uveitis animal model previously used by our group³³ and then compared to the intravitreal pharmacokinetics in healthy eyes. To induce uveitis, rats were inoculated into the right posterior paw with 1 mg/kg Escherichia coli LPS diluted in 0.1 mL phosphate-buffered saline by using a BD Micro-Fine syringe (BD, Oxford, UK) with 30-G needles. The presence of uveitis was assessed by direct inspection of the eye, using the surgical microscope. The animals were kept under such conditions for 24 hours. To reduce the number of animals, the influence of volume and presence or absence of uveitis were examined only for ¹⁸F-NaF (monoexponential kinetics) and ¹⁸F-FDG (biexponential kinetics). Four animals (eight eyes) were used in each condition studied. 

Data Acquisition and Analysis

PET Data Acquisition.

After the intravitreal injections of 1 MBq in each eye for all experimental conditions, dynamic PET acquisition was carried out to generate eight images of 15 minutes' duration for the first 1.5 hours. Afterwards, single PET images were obtained at 4 and 6 hours after drug administration. PET and CT images were acquired by using an Albira PET/CT Preclinical Imaging System (Bruker Biospin, Woodbridge, CT, USA). Animals were kept under anesthesia with a mask (1.5% isoflurane in oxygen). Respiration frequency and body temperature were monitored during the anesthesia period. The PET subsystem comprises three rings of eight compact modules based on monolithic crystals coupled to multianode photomultiplier tubes, forming an octagon with an axial field of view (FOV) of 40 mm per ring and a transaxial FOV of 80 mm in diameter. The CT system comprises a commercially available microfocus x-ray tube and a CsI scintillator 2D pixelated flat panel x-ray detector. Scatter and random coincidences were corrected by using the protocols implemented in the scanner. Attenuation correction was not performed. Images were reconstructed by using the maximum likelihood expectation maximization algorithm. Twelve iterations were performed with a reconstructed image pixel size of 0.4 × 0.4 × 0.4 mm³. 

PET Data Analysis.

After reconstruction, quantitative measurements were obtained by using the Amide's Medical Image Data Examiner.³⁵ Different regions of interest (ROIs) were manually drawn containing the signal on each eye. The ROIs were then replicated on the different temporal image frames to obtain the decrease curve of the radioisotope over time, conveniently corrected for radioactive decay. 

Statistical Analysis.

The curves of percentage of radiotracer in the eye versus time were fitted to the mono- and bicompartimental pharmacokinetic model by using nonlinear least squares regression analysis. The area under the percentage of radiotracer time curve Display Formula \(\def\upalpha{\unicode[Times]{x3B1}}\)\(\def\upbeta{\unicode[Times]{x3B2}}\)\(\def\upgamma{\unicode[Times]{x3B3}}\)\(\def\updelta{\unicode[Times]{x3B4}}\)\(\def\upvarepsilon{\unicode[Times]{x3B5}}\)\(\def\upzeta{\unicode[Times]{x3B6}}\)\(\def\upeta{\unicode[Times]{x3B7}}\)\(\def\uptheta{\unicode[Times]{x3B8}}\)\(\def\upiota{\unicode[Times]{x3B9}}\)\(\def\upkappa{\unicode[Times]{x3BA}}\)\(\def\uplambda{\unicode[Times]{x3BB}}\)\(\def\upmu{\unicode[Times]{x3BC}}\)\(\def\upnu{\unicode[Times]{x3BD}}\)\(\def\upxi{\unicode[Times]{x3BE}}\)\(\def\upomicron{\unicode[Times]{x3BF}}\)\(\def\uppi{\unicode[Times]{x3C0}}\)\(\def\uprho{\unicode[Times]{x3C1}}\)\(\def\upsigma{\unicodeTimes]{x3C3}}\)\(\def\uptau{\unicode[Times]{x3C4}}\)\(\def\upupsilon{\unicode[Times]{x3C5}}\)\(\def\upphi{\unicode[Times]{x3C6}}\)\(\def\upchi{\unicode[Times]{x3C7}}\)\(\def\uppsy{\unicode[Times]{x3C8}}\)\(\def\upomega{\unicode[Times]{x3C9}}\)\(AUC_0^{360}\) from zero to infinity was calculated by log-trapezoidal rule. The statistical analysis of experiments was performed by using a 1-way analysis of variance (ANOVA) and Tukey's multiple comparisons test. The nonlinear fitting and the statistical analysis were made by using the GraphPad Prism 6.0 software (2014; GraphPad Software, Inc., San Diego, CA, USA).  

Results

All radiolabeled molecules were clearly detected in the vitreous cavity at the initial time of the study and it was possible to observe how the signal decreased over time. Figure 2 shows the coronal views of the fused PET/CT images from the initial frame (10 minutes after the injection) to the last frame (360 minutes after the injection). 

Figure 2

 

Fused image PET/CT showing the signal evolution in the rat eyes throughout time (minutes).

Effect of the Type of Radiolabeled Molecules

The values measured from the ROI, containing each eye throughout time, were obtained for the three radiolabeled molecules, giving rise to significantly different kinetic curves (Fig. 3A). On the one hand, the clearance curves from ¹⁸F-FDG and ¹⁸F-Choline tracers appeared to fit a two-compartment model with a biphasic clearance from the vitreous. The obtained average intravitreal half-lives for these radiolabeled molecules were13.99 minutes for ¹⁸F-FDG and 35.18 minutes for ¹⁸F-Choline for the initial rapid elimination phase (α), and 214.2 minutes and 1351 minutes, respectively, for the slow elimination phase (β). Table 1 shows the pharmacokinetic parameters obtained by fitting the data to a bicompartmental model. On the other hand, the clearance curve from ¹⁸F-Na showed a one-compartment pharmacokinetic model, and the average intravitreal half-life was 113.2 minutes. Table 2 shows the pharmacokinetic parameters obtained by fitting the data to a one-compartment pharmacokinetic model. 

Figure 3

 

Influence of the drug type on its intravitreal release (mean ± SD, n = 8). (A) Intravitreal pharmacokinetic profile of ¹⁸F-FDG, ¹⁸F-NaF, and ¹⁸F-Choline after intravitreal injection of 4 μL. (B) Representation of \(AUC_0^{360}\) (% min) for all radiotracers. *1-way ANOVA analysis and Tukey multiple comparison test show significant differences among the three different compounds (α < 0.01).

Table 1

 

Pharmacokinetic Parameters Obtained by Fitting the Data to a Bicompartmental Model for ¹⁸F-FDG and ¹⁸F-Choline

 

Table 2

 

Pharmacokinetic Parameters Obtained by Fitting the Data to Monocompartmental Model With ¹⁸F-NaF

 

When comparing the area under the curve between 0 and 360 minutes (Display Formula \(AUC_0^{360}\)) among three radiolabeled molecules, it was observed that ¹⁸F-Choline remains significantly longer in the eye than ¹⁸F-FDG and ¹⁸F-NaF (Fig. 3B).  

The radiolabeled molecules leave the eye and reach the systemic circulation, following different kinetic curves. Furthermore, the distribution at system level is also significantly different. Figure 4 shows that ¹⁸F-NaF is captured by bone structures, while ¹⁸F-FDG and ¹⁸F-Choline are captured by internal organs. 

Figure 4

 

Representation of the systemic distribution of radiotracers at different times after intravitreal administration. (A) Coronal views after injection of ¹⁸F-FDG. (B) Sagittal views after injection of ¹⁸F-NaF. (C) Coronal views after injection of ¹⁸F-Choline.

The radiolabeled molecules used for the intravitreal injection had radiochemical purity for ¹⁸F-FDG higher than 95% with a specific activity of approximately 1000 MBq/mL. The ¹⁸F-Choline had radiochemical purity higher than 95% with a specific activity of approximately 500 MBq/mL. All radiotracers showed percentages of fluorine bound to the radiotracer that were higher than 95% at 8 hours post synthesis. The osmolality of all radiolabeled solutions was approximately 280 ± 10 mOsm/kg with a pH ≈7.4. 

Effect of the Injected Volumes

Figure 5 shows no differences between the different volumes of intravitreal injections (2, 4, and 7 μL) for ¹⁸F-Na and ¹⁸F-FDG radioisotopes, which follow the same kinetics as previously described in Figure 3A. Tables 1 and 2 show that no statistically significant differences were found between pharmacokinetic parameters in relation to the injected volumes of both ¹⁸F-FDG (Table 1) and ¹⁸F-Na (Table 2). Finally, it should be noted that a transient vascular collapse in the retinal vessels was observed after administration of 7 μL, but not for 2 and 4 μL. 

Figure 5

 

Influence of the injection volume on vitreal release (mean ± SD, n = 8). Intravitreal pharmacokinetic profile of ¹⁸F-FDG (A) and ¹⁸F-NaF (B) after intravitreal injection of 2, 4, and 7 μL.

Effect of the Presence of Inflammation

Figure 6A shows that inflammation slightly, but with statistical significance, increased the vitreous clearance of ¹⁸F-FDG. This effect was quantified by comparing the Display Formula \(AUC_0^{360}\) of radiolabeled molecules in uveitis and under normal conditions. Figure 6B shows that eyes with uveitis had smaller Display Formula \(AUC_0^{360}\) than normal eyes. In addition, statistically significant differences were found between the pharmacokinetic parameters in uveitis and normal conditions for the case of ¹⁸F-FDG (Table 1). It must be mentioned that animals receiving an LPS injection developed a fibrinous reaction in the anterior chamber of the eye, which produced a pupillary membrane and an irregular pupil after drug-induced mydriasis, caused by the adhesion of the iris to the lens (Fig. 7). The uveitis model was successfully achieved in the same way as obtained in our previous studies.³⁶  

Figure 6

 

Influence of inflammation on vitreal release (mean ± SD, n = 8). (A) Intravitreal pharmacokinetic profile of ¹⁸F-FDG and ¹⁸F-NaF after a 7-μL intravitreal injection in normal eyes and in eyes with uveitis. (B) Representation of \(AUC_0^{360}\) (% min) for ¹⁸F-FDG and ¹⁸F-NaF in these conditions. *Statistical significant differences between normal and uveitis eyes for α < 0.01.

Figure 7

 

Anterior segment of two eyes 24 hours after pad injection of LPS, showing signs of uveitis. Left: Fibrinous reaction producing a pupillary membrane. Right: Irregular pupil after drug-induced mydriasis caused by the adhesion of the iris to the lens.

Discussion

Intravitreal injections are increasingly used in a multitude of retinal ophthalmic conditions such as age-related macular degeneration,³⁷ diabetic macular edema,³⁸ macular holes,³⁹ branch and central retinal vein occlusion,⁴⁰ and endophthalmitis.⁴¹ The development of new intravitreal drugs or systems that modify their release involves wide preclinical development⁴² in which pharmacokinetic studies play a key role.⁴³ 

The use of small animals, such as Sprague Dawley rats, has many advantages because of their small size, the availability of research animal facilities, and multiple disease models suitable for them.^44,45 However, since they have a small vitreous volume, classic pharmacokinetic studies become difficult, with in vivo imaging being an ideal technique, as no invasive modalities are required to obtain experimental results.^46,47 To the best of our knowledge, our work is the first study of intravitreal pharmacokinetics with PET/CT in rats. Previous intravitreal pharmacokinetic studies have required larger numbers of animals and more complex techniques to determine vitreous drug levels at different time points.^48–50 In our study serial measurements were obtained at multiple time points after the intravitreal injection in the same animal. The advantage of preclinical PET/CT images in this field is very important because the technique is noninvasive, and it yields images in 3D and real time.⁵¹ PET/CT is also becoming a relevant procedure for ophthalmic research, as it has been used for diagnosis of intraocular tumors,⁵² neurophysiological studies,^53,54 or pharmacokinetic studies with topical ophthalmic formulations.²³ Although PET is a very sensitive technique, it presents some limitations related to low spatial resolution. As an example, the delineation of the vitreous area is troublesome and challenging owing to the small size of the eyeball, and therefore our measurements cannot be restricted exclusively to the vitreous area. 

More than 10% of currently used drugs contain fluorine atoms that can be labeled with ¹⁸F. Moreover, the substitution of oxygen atoms or hydroxyl groups by fluorine is relatively easy with no critical changes in the properties of the molecule.⁵⁵ Fluoride and oxygen have a very similar radius, whereas that of hydrogen is slightly smaller (van der Waals radii are 1.47, 1.57, and 1.20 angstroms, respectively); therefore, changing oxygen or hydrogen for fluoride does not entail substantial modifications in the molecular structure by steric impediments. Furthermore, in terms of Taft E_s parameters,⁵⁶ fluoride and hydroxyl substituents have very similar characteristics (+0.78 vs. +0.69); therefore, their substitution does not compromise either the structural activity of the compound or its interaction with receptors. The electronegativity of fluoride and hydrogen atoms is different (4.0 vs. 2.1), hence interchanging them can substantially affect the physicochemical properties of the molecule (pKa, hydrogen bond capacity, or lipophilicity). On the contrary, fluoride and oxygen have similar values (4.0 vs. 3.5), so no major changes should be expected when interchanged.⁵⁵ Owing to the relatively short half-life of ¹⁸F, the fluorinated radiotracers have limited use in studies of pharmacokinetics or biodistribution of drugs with long half-lives in the vitreous cavity. For these long-term studies, using other radiotracers with long half-lifes such as ¹²⁴I (Kuntner et al.⁵⁷ and Dangl et al.⁵⁸) or ⁸⁹Zr (Van Loon et al.⁵⁹) is more adequate. 

Fluorinated radiotracers, as the ones used in this work, have the advantage of their low positron emission energy (the lowest of all the radiolabels used in PET). Furthermore, the greater sensitivity of modern PET technology allows the use of low radioactivity levels, so the dose received and absorbed by the animal is significantly below the dose limit.⁶⁰ Additionally, during the disintegration of ¹⁸F, no γ rays or α and β particles are emitted, reducing the dose received by animals and increasing safety.⁶¹ On the other hand, cytotoxicity and acute irritation of fluorinated radiotracers have been described as safe in previous reports.²³ In our study, no alterations in the eye of the animals were observed after the administration of the fluorinated radiotracers. 

Our findings showed significant differences between the different radiolabeled molecules we used. The reason for these differences could rest on the mechanism used for crossing the BRB. In the rat retina there are transporters for glucose and cationic amino acids, which probably are used by ¹⁸F-FDG and ¹⁸F-Choline to leave the vitreous cavity.⁶² The biexponential kinetics we observed is also common for intravitreal drugs such as bevacizumab and ranibizumab.^50,63 Furthermore, it must be mentioned that hyaluronic acid, which is part of the vitreous humor, has a highly negative charge at physiological pH levels. Because of this, it could interact with positively charged molecules, such as choline, by generating polyelectrolyte complexes with low solubility.⁶⁴ This is probably the reason why choline is released at a slower rate than glucose. On the other hand, our findings showed that ¹⁸F-NaF is eliminated from the vitreous, following monoexponential kinetics, which could be explained by assuming passive diffusion through the BRB because this compound diffuses freely across membranes.⁶⁵ It would be similar to the release kinetics of other intravitreal drugs, such as aflibercept.^66,67 

Our findings showed that the injected volume had no significant influence on vitreous drug clearance. Different studies have been carried out in human eyes⁶⁸ and in murine models^69,70 using a wide range of intravitreal injection volumes (2–20 μL), but they did not include an evaluation of their impact on the vitreous drug clearance. On the other hand, it has been pointed out that an increase of intraocular pressure could result in an increase of hydraulic flow, derived from the excess of volume introduced.¹⁰ This increase in intraocular pressure could be the cause of the transitory collapse we observed with the administration of 7 μL. However, this process seems not to have an effect on the vitreous clearance of low-molecular-weight drugs,^71,72 such as the ones we used, where all radiolabeled molecules had molecular weights below 500 Da. Finally, it has to be mentioned that the vitreous volume of a rat is smaller than that of humans (approximately 50 μL in rats versus 4.5 mL in humans)⁷³ (Vezina M, et al. IOVS 2011;52:ARVO E-Abstract 3219). This difference must be kept in mind if our results are to be translated to humans. 

Our results showed an increase in the intravitreal clearance of the ¹⁸F-FDG radiotracer in eyes with inflammation (uveitis) when compared to healthy eyes. On the contrary, no significant differences were observed for ¹⁸F-NaF. Studies using MRI techniques have shown that inflammation in rabbit eyes, induced by LPS, can increase the permeability of BRB.^74,75 On the other hand, additional studies have demonstrated that in inflammatory conditions, as in tumors, a high FDG uptake and a high GLUT-1 expression level is observed.⁷⁶ Of note, ¹⁸F-NaF is not affected by changes produced by the inflammatory process probably because it is freely diffusible across membranes.⁶⁵ However, the increase in permeability and GLUT transporter under inflammatory conditions can increase the clearance of ¹⁸F-FDG from the vitreous. Since the magnitude of the clearance changes we found were small, it would be necessary to carry out additional studies to properly establish the influence of inflammation on the BRB permeability. It is possible that the severity of the inflammatory process determines the increase of BRB permeability and hence, the intravitreal clearance rate. 

Finally, although the effect of inhaled anesthesia, in particular isoflurane, on drug permeability has been extensively studied in the blood–brain barrier (BBB), no studies have shown any type of modification in the status of the BRB.⁷⁷ Inhaled isoflurane in rats decreases the transfer of small hydrophilic molecules across the BBB, either by reducing the perfused capillary surface area or by a direct effect of isoflurane on the permeability of the BBB.⁷⁸ 

Even though BRB and BBB are thought to have similar properties owing to their similar anatomic features, some differences in the capillary endothelia at BBB and BRB have been found.⁷⁹ These differences could cause a variation in the permeability across BRB compared to BBB,⁸⁰ which would ultimately affect the half-life and clearance of drugs injected into the vitreous. Therefore, we cannot assume that the observed increase in permeability through BBB would lead to a similar effect in BRB. Studies about the effect on inhaled anesthesia on BRB permeability would be of great interest. 

Conclusions

In summary, we demonstrated for the first time the usefulness of a PET-based methodology for the study of different factors influencing intravitreal pharmacokinetics in rats. This can be a powerful tool to develop new drugs aimed at treating ocular conditions, using intravitreal administration. 

Acknowledgments

Supported by Instituto de Salud Carlos III Rio Hortega Research Grant CM15/00188 (AF-F) and Miguel Servet Research Grant CP12/03162 (JP-M), MICINN Ramon y Cajal RYC-2015-17430 (PA), and partially supported by the government research funding agencies of Xunta de Galicia, Spain (ED431C-2016-025, ED341D-R2016-032), and by ISCIII (RETICS Oftared, RD16/0008/0003 and RD12/0034/0017) cofunded by FEDER. 

Disclosure: A. Fernández-Ferreiro, None; A. Luaces-Rodríguez, None; P. Aguiar, None; J. Pardo-Montero, None; M. González-Barcia, None; L. García-Varela, None; M. Herranz, None; J. Silva-Rodríguez, None; M. Gil-Martínez, None; M.A. Bermúdez, None; A. Vieites-Prado, None; J. Blanco-Méndez, None; M.J. Lamas, None; F. Gómez-Ulla, None; Á. Ruibal, None; F.J. Otero-Espinar, None; F. González, None