6.5  Clearance Pathways from the Vitreous Compartment

phosphatidylethanolamine polyethylene glycol (DSPE-PEG) to systems which are smaller than 500 nm, no binding of the coated lipoplexes to vitreal collagen strands was observed above a content of 16.7 mol% and a clear relationship between the DSPE-PEG concentration and aggregation could be observed in the micrographs.

6.5.2  Aqueous Clearance

Friedrich’s group has shown that injection position and volume has significant influence on clearance kinetics of model compounds fluorescein and fluorescein glucuronide (Friedrich et al. 1997a, b). Different injection sites (including behind the lens, at the hyaloid membrane, central injection and injection next to the retina) were found to influence the measured retinal permeability of fluorescein from 1.94 × 10−5 up to 3.50 × 10−5 cm/s. Thus the site of the intravitreal injection of fluorescein is predicted to influence distribution and permeability through the retina. In addition, it was calculated that the mean concentration remaining in the vitreous at 24 h varied up to a factor of 3.8-fold dependent on initial location of the smaller volume of 15 mL. It was also shown that increasing the volume from 15 to 100 mL reduced the magnitude of these changes to approximately 2.5-fold at 24 h (Friedrich et al. 1997a).

The rapid turnover of aqueous humour in the anterior chamber is the main motive force for forward clearance. All compounds injected intravitreally can be removed through this bulk flow system. The majority of materials can effortlessly move across the hyaloid membrane; the central anterior position of the lens being the main barrier to this forward movement (Xu et al. 2000; Worst and Los 1995). Thompson and Glaser showed that the flux of 20 and 70 kDa dextran from the vitreous into the anterior chamber increased significantly after extracapsular lensectomy with posterior capsulotomy (Thompson and Glaser 1984). In addition, according to the study performed by Stepanova et al., the transport mechanism present at the lens epithelium generates uni-directional flow that moves fluids towards the retina rather than the anterior chamber (Stepanova et al. 2005). Therefore, materials tend to move around the edge of the lens, instead of diffusing across the highly packed 20 nm collagen meshwork (Worst and Los 1995). Substances that successfully enter the anterior chamber are subsequently removed along with aqueous humour by the trabecular and uveoscleral outflow (Cunha-Vaz 1997).

The aqueous drainage at the anterior chamber generates a sustained “sink condition” for intravitreally administered substances, resulting in the formation of a concentration gradient, originating from the injection pocket, that spreads across the vitreous cavity. Maurice illustrated a clearance process parallel to the posterior capsule of the lens with the lowest concentration located at the hyaloid membrane gradually increasing towards the retina (Araie and Maurice 1991).

Typically, hydrophilic and larger molecules that are not able to exit through the retina are removed via the anterior route. Atluri and Mitra investigated the vitreal disposition of short-chain aliphatic alcohols with varying degrees of lipophilicity in

rabbits using ocular microdialysis techniques. Their findings reveal that methanol achieved the highest concentration at the anterior chamber, whereas the concentration of the more lipophilic 1-heptanol was undetectable (Atluri and Mitra 2003). Araie and Maurice have also shown that, as opposed to fluorescein which is smaller and lipophilic, larger and hydrophilic molecules such as fluorescein glucuronide and fluorescein isothiocyanate dextran have poor retinal penetration. A steep concentration gradient from the vitreous to the posterior chamber was noted and flux through the retina was small indicating loss through the aqueous route (Araie and Maurice 1991). In addition, a small human study utilising samples taken from the aqueous humour, demonstrated that the anterior elimination pathway is important for intravitreal clearance of the steroid, triamcinolone (Beer et al. 2003).

If we exclude transporter effects and phagocytosis, the retinal barrier of the inner eye should be problematic for large molecules with low flux which would to be removed by the aqueous system (Atluri and Mitra 2003; Urtti 2006). A relatively small biopharmaceutical, an oligonucleotide with an approximate molecular weight of 7 kDa, exhibited rapid anterior clearance following intravitreal injection, with less than 7% remaining after 7 days (Dvorhik and Marquis 2000). A study of a larger sized biopharmaceutical, rituximab, in rabbits, suggested that clearance of rituximab occurred via the aqueous route. It was suggested that Rituximab diffuses through the vitreous, between the lens and ciliary body, into the anterior chamber for removal (Kim et al. 2006). Similarly, removal via aqueous humour is thought to represent the predominant clearance pathway of bevacizumb, following intravitreal administration in man (Krohne et al. 2008). Bakri et al. described the clearance kinetics of intravitreal bevacizumab in Dutch-belted rabbits using a non-compartmental model and concluded that bevacizumab was cleared through the anterior pathway with an estimated intravitreal half-life of 4.32 days (Bakri et al. 2007).

6.5.3  Retinal Clearance

The posterior elimination pathway has been proposed to be the primary route for small and lipophilic molecules. Once removed from the retina, materials will be subsequently transported away by the choroidal blood flow. If melanin binding is significant, accumulation in the melanocytes of the uveal tract will occur. The RPE is able to remove material by passive diffusion through the paracellular and/or transcellular routes. In vitro, the retinal permeability is 8–20 times higher for lipophilic than hydrophilic molecules, suggesting a higher efficiency of the transcellular pathway (Pitkänen et al. 2005).

Inflammation of the RPE, encountered in patients with endophthalmitis, damages retinal pump function thereby decreasing the intravitreal half-life of molecules eliminated by this system (Ficker et al. 1990). The RPE therefore forms an important component of the blood-retinal barrier and contains retinal glial cells

and the endothelium of retinal blood vessels (Cunha-Vaz 1997). An early study by Mosley (1981) modelled the movement of [133Xe] xenon from the vitreous through the retina, using samples taken from the vortex vein and concluded that the radioisotope was cleared from the vitreous, to the retina and into the choroid circulation with a mean transit time of 27 min. A pharmacokinetic model of data from a rabbit study suggested the antiviral used in the treatment of cytomegalovirus retinitis, ganciclovir, is eliminated across the retinal surface (Tojo et al. 1999). In addition, following intravitreal administration of memantine, high concentrations were found in the choroid and RPE, suggesting posterior elimination (Koeberle et al. 2003).

The size of antibody fragments begins to approach nanoparticulate dimensions and therefore data from nanoparticulate movement might be a useful predictor of large anti-VEGF agents. Sakurai et al. showed that particles of sizes 200 nm and below can transverse the retina but 2 mm particles were found to mainly clear through the trabecular meshwork (Sakurai et al. 2001). Pitkänen et al. have also reported that the permeability of carboxyfluorescein (376 Da) was 35 times higher as compared to FITC-dextran 80 kDa (Pitkänen et al. 2005). In contrast, data obtained by Dias and Mitra showed that FITC-dextran at a molecular weight of 38.9 kDa was predominantly removed from the vitreous through the retina, an observation attributed to the possible presence of channel-mediated transport mechanism at the retina for large and hydrophilic molecules (Dias and Mitra 2000). A comparison between the characteristics of forward and retinal clearance is illustrated in Table 6.1.

Table 6.1  Comparison between the forward and retinal routes of clearance