The size and surface characteristics of NDDS has long been recognized as important parameters, influencing how the NDDS will distribute after intravitreal injection. Equally, the vitreous network has been proposed as a barrier in itself.
18,26 The size of the nanocarrier needs to be small enough that it can move through the pores in the vitreous, reported to be between 1 µm and 500 nm
17 in the bovine vitreous, which is considered a good model of the human vitreous.
27–29 Particles larger than 500 nm have been reported to be trapped or show very limited diffusion in the vitreous
17 and most reports involving nanocarriers for intravitreal injection have diameters <300 nm.
14 However, reposts suggest that small nanocarriers (<25 nm) are cleared faster than larger nanoparticles, (>200 nm) and most likely through the anterior route.
30
Another important physicochemical property of nanocarriers is surface chemistry and surface charge. The vitreous is reported to have an overall negative charge
18 and particles with strong positive surface charges have been reported to become electrostatically trapped in the hydrogel network.
17,18,28,31–35 We
16 and others
20 have shown that PEGylated liposomes with moderate positive surface charges can move in the vitreous, although more slowly than their negatively charged counter parts, and transfection studies with nanoparticles performed in rats, suggest that a positive charge can be beneficial for retinal uptake of particles and transfection of retinal cells.
36,37 This is supported by the recent report by Christensen et al. who showed that cationic PEGylated liposomes adsorbed better to the retina in ex vivo porcine eye than anionic PEGylated liposomes.
38 Here, we tested PEGylated liposomes that were 128.0 ± 8 nm and 120.2 ± 2 nm, hence in a size range that can be expected to diffuse in the vitreous, while decreasing the clearance rate through the anterior segment. We did not observe a significant difference in vitreous and retina distribution of both anionic and cationic liposomes. The lack of an effect of surface charge can likely be attributed to the very limited vitreous in mice, which is approximately 1000 times smaller than in humans.
39,40 At the 24-hour timepoint, this biodistribution is altered (albeit as a trend rather than significant) and we observe a higher Pt concentration in the vitreous from the cationic liposomes compared with anionic liposomes and free OxPt. This likely reflects a slower diffusion, or more pronounced trapping, of the cationic liposomes in the negatively charged central vitreous hydrogel.
We tested the liposome stability in vitro in respect to their OxPt release in buffer and found that the liposomes only leak up to approximately 2.5% of their OxPt cargo at 37°C. Liposomal stability, however, might be affected by the proteins and salts present in the vitreous. Our previous study on similar liposomes, using fluorescent correlation spectroscopy (FCS) of both a hydrophobic membrane bound fluorophore and a hydrophilic fluorophore in the liposome core, showed that the liposomes remain intact when diffusing in the porcine vitreous,
16 and we therefore expect that the liposomes will retain their OxPt cargo when diffusing through the vitreous.
The incorporation of cationic lipids has been reported to increase the cytotoxicity of liposomes.
41,42 Here, we compared the cytotoxicity of the two liposome formulations in two retinal cells line (mREC and ARPE-19) and found no significant difference in the cytotoxic response, even though the ARPE-19 cells had an almost 250 times higher uptake of the liposomal Pt, compared to the mREC. No toxic response was observed in the cells within the 24-hour timeframe. The lack of cytotoxicity is likely explained by the mechanism of action of OxPt, involving DNA damage and blockage of DNA repair,
43 as these effects will mainly be observed in relation to cell division, not likely to happen within 24 hours. The absence of cytotoxicity at the 24-hour time point is important in ensuring that uptake and cellular transport of the liposomes is performed by healthy cells in vivo. Similar cationic liposomes were reported by Christensen et al. to be well tolerated by the retina.
38
We compared the distribution of intravitreally injected free OxPt and liposomal OxPt at 2 hours and 24 hours. In line with previous literature, we observed that the small molecule free OxPt cleared faster than the liposomal OxPt,
44 showing that even in the small volume vitreous of mice, a positive effect of enhanced retention of nanocarrier encapsulated small molecules can be observed. Interestingly, despite a higher retention in the vitreous, the concentration in the retina remained relatively low after intravitreal injection with liposomes, whereas the concentration in the RPE-sclera fraction increased simultaneously with clearing from the vitreous. The intercellular spacing in the retina is estimated to be <20 nm,
45 hence smaller than the liposomal diameter. This suggests that active transport mechanisms may be involved in the trans-retinal clearing of liposomes, whereas free OxPt being significantly smaller, could possibly diffuse passively through this route.
When active uptake was inhibited by pre-injection of chloroquine, we observed a significant decrease in the OxPt concentration in the RPE-sclera and a slight increase in other tissues, for example, the vitreous and cornea. The malaria drug chloroquine is a broad inhibitor of clathrin-mediated endocytosis, recognized as the most important cellular uptake mechanism of nanocarriers.
25,46 Additionally, chloroquine inhibits endosome trafficking, disrupting intra-cellular transport by increasing lysosomal pH,
24 and disorganizing the golgi apparatus.
47 High concentrations of chloroquine have been found to be toxic to photoreceptors after intravitreal injection in cats.
48 However, no effect was observed on the inner limiting membrane (ILM) or Müller glial cells.
49 The low concentrations used in this work is not expected to result in any retinal toxicity.
50,51
We propose that clathrin mediated uptake and endosomal trafficking is involved in the transport of liposomes from the posterior vitreous. It has been argued that NDDSs are too large to penetrate through the finer protein network of the ILM,
29 an effect that might be less pronounced in the rodent eye, because of the thinner ILM in mice compared to humans and larger animals.
52,53 An important component on the ILM, however, is the end-feet of the Müller glial cells, that span the whole length of the retina from the ILM to the RPE.
20,54 Koo et al. showed that fluorescently labeled polymeric nanoparticles found, within retinal layers, were located inside the cell bodies of Müller cells.
55 Müller glial cells are phagocytotic cells,
56–58 and are the most abundant glia cells in the retina and act as the main transport cell in the retina (transporting nutrients, ions, water, and waste products to ensure retinal homeostasis).
59 Müller glial cells can also phagocytose material larger than 500 nm, transporting it to the outer retina.
14 Interestingly, Müller glial cells in the rabbit were found to be able to penetrate the ILM and endocytose intravitreally injected carbon nanoparticles with a diameter of 20 nm,
60 and large (1 µm diameter) egg-lecithin coated silicone particles were also found to be phagocytosed by the Müller glial cells.
61 The chloroquine data presented here indicate that our liposomes are most likely transported through an endocytosis/transcytosis pathway rather than by phagocytosis.
Glial cells are enriched in secretory lysosomes,
62 and we speculate that it is via this route (liposome endocytosis followed by secretion) that liposomal Pt is transported to the RPE-sclera. If this is the case, then this may be a major clearing mechanism for NDDS in the eyes, but further experiments are needed to conclude this. However, Müller glial cell mediated transport of nanocarriers to the RPE-sclera may provide a route to target diseases associated with RPE breakdown, such as AMD. The observed active transport of the Pt loaded liposomes highlights the need for in vivo evaluations of NDDS intended for intravitreal injection. Whereas it is recognized that the transport dynamics of NDDS are likely slower in the human eye compared with the mouse eye as a result of the size difference, many promising therapies (e.g. cell transplantations
10 and gene-replacement therapies)
63,64 are first tested in mice and a better understanding of the active transport mechanisms of NDDS in the murine retina can aid the development of these novel treatments.