|Home | About | Journals | Submit | Contact Us | Français|
Choroidal neovascularization (CNV) leads to loss of vision in age-related macular degeneration (AMD), the leading cause of blindness in adult population over 50 years old. In this study, we developed intravenously administered, nanoparticulate, targeted nonviral retinal gene delivery systems for the management of CNV. CNV was induced in Brown Norway rats using a 532 nm laser. We engineered transferrin, arginine–glycine–aspartic acid (RGD) peptide or dual-functionalized poly-(lactide-co-glycolide) nanoparticles to target delivery of anti-vascular endothelial growth factor (VEGF) intraceptor plasmid to CNV lesions. Anti-VEGF intraceptor is the only intracellularly acting VEGF inhibitory modality. The results of the study show that nanoparticles allow targeted delivery to the neovascular eye but not the control eye on intravenous administration. Functionalizing the nanoparticle surface with transferrin, a linear RGD peptide or both increased the retinal delivery of nanoparticles and subsequently the intraceptor gene expression in retinal vascular endothelial cells, photoreceptor outer segments and retinal pigment epithelial cells when compared to nonfunctionalized nanoparticles. Most significantly, the CNV areas were significantly smaller in rats treated with functionalized nanoparticles as compared to the ones treated with vehicle or nonfunctionalized nanoparticles. Thus, surface-functionalized nanoparticles allow targeted gene delivery to the neovascular eye on intravenous administration and inhibit the progression of laser-induced CNV in a rodent model.
The purpose of this study was to develop systemically administrable nanoparticle-based targeted delivery systems for anti-vascular endothelial growth factor (VEGF) intraceptor gene delivery to the posterior segment of the eye for the management of choroidal neovascularization (CNV), the cause of blindness due to wet age-related macular degeneration (AMD). AMD is the leading cause of blindness in population above 50 years of age.1 Because age is the major risk factor for AMD, with increasing longevity the incidence is increasing. AMD is characterized by the accumulation of membranous debris on both sides of the retinal pigment epithelial (RPE) basement membrane. Pathological presentation includes drusen, basal deposits, atrophy of RPE and choriocapillaris, RPE detachment and CNV. CNV in the macula is the major cause of central vision loss, especially when it is located subfoveally. Thus, the mainstay of AMD treatment is aimed at inhibiting or destroying CNV.
The therapies for CNV can be categorized into destructive therapy and pathway-based therapy.2 Destructive therapy includes laser photocoagulation and photodynamic therapy. Laser photocoagulation is applicable to a relatively small fraction of AMD patients with well-defined extrafoveal CNVs. Furthermore, the technique suffers from a high rate of recurrence of subfoveal CNV in ~50% of patients within the first 5 years after surgery. Although photodynamic therapy using verteporfin is applicable to the treatment of subfoveal CNV, it is palliative in nature and there is up to 5% chance of >4 line visual loss within 7 days of treatment.
Pathway-based therapy mainly aims at inhibiting the process of angiogenesis by inhibiting a proangiogenic factor or stimulating an antiangiogenic factor.3 The most investigated and clinically efficacious approaches in this therapeutic category aim at inhibiting VEGF-A. Two products, Macugen and Lucentis based on this approach have been approved by Food and Drug Administration (FDA) and are being marketed for the treatment of AMD. Macugen does not bind all the isoforms of VEGF-A. This might be one of the reasons contributing to its limited clinical success. Other potential reasons include the nature of the formulation or its availability at the target cells and tissues. Lucentis binds all the VEGF isoforms, but needs to be delivered by intravitreal injection. Although, there is wide consensus on the efficacy of VEGF inhibition for AMD treatment, a safer and less invasive delivery system is yet to be designed to maximize the therapeutic benefits of these modalities. In the current work, we have investigated the delivery and efficacy of anti-VEGF intraceptor plasmid loaded in nonviral gene delivery system upon intravenous administration.
Among various VEGF inhibitory strategies, anti-VEGF intraceptor approach stands out as being the only approach capable of targeting VEGF intracellularly.4,5 Intracellular VEGF targeting has an advantage in case of cells responding to their own VEGF secretion in an autocrine fashion. The autocrine loops could be external or internal.6 Internal autocrine loops are refractory to inhibitors that fail to penetrate the intracellular compartment.7 Existence of internal autocrine loops has been reported in some cell types, such as hematopoietic stem cells7 and leukemic cells.8 Contrary to the traditional view of autocrine signaling, in an internal autocrine loop, the mitogenic signal of VEGF is transduced intracellularly without factor secretion. For such an intracellular signal transduction to happen, VEGF receptors need to be present intracellularly. Indeed, in leukemic cells, VEGF receptor 2 has been reported to be predominantly expressed in cell nuclei.8 Further, this internal autocrine signaling could be inhibited only by an intracellular KDR inhibitor but not by an extracellularly acting VEGF-neutralizing antibody. Autocrine loops have also been reported in endothelial cells.9 Anti-VEGF intraceptor, Flt23K, is a recombinant construct of VEGF-binding domains 2 and 3 of VEGFR-1/Flt-1 receptor coupled with the endoplasmic reticulum (ER) retention signaling sequence Lys-Asp-Glu-Leu (KDEL). KDEL is a peptide retention signal that binds ER retention receptors and thereby prevents secretion of endogenous ER proteins coupled to KDEL. Anti-VEGF intraceptor binds VEGF intracellulary and sequesters it within the ER. Thus, intraceptors, in addition to inhibiting VEGF secretion, might interfere with the internal autocrine loops.
Gene delivery to the eye is being extensively investigated. Various reasons make gene delivery to eye a promising therapeutic approach. First, the biochemical changes involved in vision-threatening ocular disorders are fairly well known. Second, the eye is relatively immune privileged. Third, local injections enable extensive use of viral vectors. Fourth, gene delivery has the potential to alleviate the cause of a disease. Genes are most commonly delivered to the posterior segment of the eye by injecting viral vectors locally into the vitreous or the subretinal space.10–12 In addition to being invasive to the eye, recent reports on biodistribution of intraocularly administered viral vectors suggest that these vectors gain access to the brain along the visual pathway, further raising concerns about the safety of locally administered viral vectors.13 Recent reports have also documented barrier nature of vitreous14,15 and neural retina16 for gene transfer to RPE on intravitreal injection. Therefore, there is a need to develop alternative gene delivery approaches for the eye potentially using safer nonviral vectors.
The application of nonviral systems for gene delivery is being increasingly advocated because of their low immunogenicity, unlimited payload capacity, absence of endogenous virus recombination, low production cost and reproducibility. A plethora of nonviral systems for intravitreal injection into the eye have been developed. These include poly-(lactide-co-glycolide) (PLGA) nanoparticles,17 albumin,18 dendrimer,19 polyethyleneglycol-substituted lysine peptides20 and liposomes.21 We and our collaborators have recently shown that Flt23K-plasmid-loaded nanoparticles sustain delivery of the plasmid for up to 5 weeks and reduce injury-induced corneal neovascularization by ~40% on intrastromal injection in BALB/c mice.4 Furthermore, the formulation was found to be nontoxic to the cornea. We have also shown using an ex vivo bovine eye model that surface functionalization increases nanoparticle delivery to the cornea on topical dosing to up to 16% as compared to just 2.4% with nonfunctionalized nanoparticles.22 Intravenous administration of functionalized PLA nanoparticles has been shown to deliver genes to brain.23 A few attempts have been made to deliver genes to the eye by intravenous route of administration. Although, a diffuse expression of gene in inner retina, RPE, iris as well as conjunctival epithelium can be achieved on intravenous administration of liposomes,24 for targeting gene expression to a specific tissue in the eye, the plasmid needs to be modified to include an eye-specific promoter.25 Moreover, to the best of our knowledge, there are no published reports demonstrating the therapeutic efficacy of intravenously administered gene delivery systems for the management of ocular disorders. Thus, there is a need to design delivery systems, which can target the gene delivery to the pathological tissue within the eye on intravenous administration and prove the efficacy of these systems to inhibit ocular disorders. This article attempts to fill this void in scientific understanding by studying the feasibility and efficacy of targeted delivery of Flt23K plasmid to laser-induced CNV lesions using surface-functionalized PLGA nanoparticles by intravenous route of administration.
To facilitate endocytosis of intraceptor-loaded nanoparticles, we functionalized the nanoparticle surface with a linear arginine–glycine–aspartic acid (RGD) peptide (GRGDSPK), transferrin or a combination of both (Figure 1). Peptides containing RGD motif can bind integrin receptors.26 Specifically, the peptide GRGDSPK binds integrin αvβ3 receptors and is thus commonly used in purification of the same.27 Integrin receptors are upregulated in ocular neovascularization. Specifically, integrin αvβ3 overexpression is observed in ocular tissues from AMD patients, whereas both integrins αvβ3 and αvβ5 are observed in vascular tissues from proliferative diabetic retinopathy patients.28 Transferrin is an 80 kDa protein that participates in the transport of iron in circulation. Iron-laden transferrin is actively taken up into retinal cells by transferrin receptor-mediated endocytosis. Increased accumulation of iron29 and enhanced transferrin levels30 have been observed in AMD retinas. Although there is no conclusive report on the expression of transferrin receptors in AMD retinas, increased iron load in the AMD retina could be because of overexpression of transferrin receptors and increased transport of iron. Thus, we hypothesized that surface functionalization of Flt23K-plasmid-loaded nanoparticles with RGD peptide, transferrin or both will further enhance gene delivery to the neovascular eye on intravenous administration.
Physicochemical properties of nanoparticles are summarized in Table 1. The effective diameter of nanoparticles ranged from ~270 to 420 nm. The size of RGD-peptide-conjugated nanoparticles increased marginally by 7 nm whereas that of transferrin and dual (RGD peptide and transferrin)-conjugated nanoparticles increased by 45 nm, as compared to unconjugated Flt23K-plasmid-loaded Nile red PLGA nanoparticles. The polydispersity index, a measure of nanoparticle size distribution, was ≤0.25 for all the groups, suggesting fairly unimodal particle size distribution for both functionalized as well as nonfunctionalized nanoparticles. All the particles had a negative ζ-potential as expected for PLGA nanoparticles.31–33 Further, anti-VEGF intraceptor plasmid loading in PLGA nanoparticles was found to be 10.11 µg mg−1 nanoparticles.
We did not observe any significant aggregation of any of the nanoparticles on incubation with phosphate-buffered saline (PBS, pH 7.4) or serum for 24 or 48 h. Also, there was no increase in the size of any of the nanoparticles on incubation in PBS for up to 48 h. This is consistent with a previous study by Zweers et al.,34 which indicated no aggregation of PLGA nanoparticles on incubation in PBS at the end of 2 days. However, we observed a marginal increase in the size of nanoparticles on incubation with serum. The size increase at 24 and 48 h was similar. On incubation of nanoparticles with serum for 48 h, the effective diameter of Flt23K-NP, Tf-Flt23K-NP, RGD-Flt23K-NP and RGD-Tf-Flt23K-NP increased from 352 to 376, 403 to 451, 359 to 378 and 414 to 463 nm, respectively. This is expected as a result of opsonization.35 An increase in PLGA nanoparticle size on incubation with plasma has been previously reported by Yang et al.35 Our size distribution plots did not indicate any aggregates in either PBS or serum incubation studies.
Nanoparticles loaded with Flt23K plasmid and Nile red could be detected only in the laser-treated (right) eye, on confocal microscopy of the posterior segment flatmount (Figure 2a). No particles could be detected on confocal microscopic examination of similarly prepared flatmounts of control (left) eye. The observation of the particles being delivered only to the laser-treated eye was consistently made in all the groups of the study (Figure 2b). This suggests that leakiness of the ocular blood vessels is important in the delivery of nanoparticles to eye on intravenous administration.
On confocal microscopic examination of posterior segment flatmounts (retina–choroid–sclera), it was observed that the retinal delivery of RGD peptide, transferrin and dual-functionalized nanoparticles was higher than that of nonfunctionalized PLGA nanoparticles (Figure 3). On quantification of relative red fluorescence intensity in each eye, the retinal delivery by various nanoparticles exhibited the following trend: RGD-conjugated nanoparticles > transferrin-conjugated nanoparticles ~ dual-conjugated nanoparticles > unconjugated nanoparticles (Figure 4). The distribution of nanoparticles within the retina was found to depend on the functionalization moiety. RGD- and dual-conjugated nanoparticles showed significant targeting to retinal blood vessels, whereas no such targeting was evident for transferrin-conjugated nanoparticles (Figure 3). Vascular targeting by RGD peptide can be explained by abundant expression of integrin receptors on vascular endothelial cells.36 Confocal microscopy under multitrack mode using red (Nile-red-loaded nanoparticles), green (green fluorescent protein; GFP) and blue (cell nuclei stained with 4′ ,6-diamidino-2-phenylindole, DAPI) revealed GFP expression at 24 h after intravenous administration of nanoparticles. The green fluorescence due to GFP colocalized with the red fluorescence of nanoparticles, suggesting that the nanoparticles could not only reach retina, but also successfully delivered the loaded plasmid.
The delivery of functionalized nanoparticles to all the tissues tested was either less than or comparable to nonfunctionalized nanoparticles (Figure 5). Nanoparticles were not detected in the brain of any animal. The delivery of nanoparticles to the remaining tissues followed the order: liver > lung > kidney ~ spleen > heart. Thus, enhancement in delivery on surface functionalization of nanoparticles with transferrin, RGD peptide or both was specifically seen in the neovascular eye.
Retinal pigment epithelial cells are one of the major sources of VEGF secretion in the retina. Thus, delivery of anti-VEGF intraceptor plasmid to RPE cells is essential for antiangiogenic efficacy in vivo. On confocal examination of ocular sections obtained at 48 h after intravenous administration of nanoparticles the GFP expression was observed in photoreceptor cell layer as well as RPE cells in case of RGD, transferrin and dual-functionalized nanoparticles (Figure 6). On the other hand, in case of nonfunctionalized nanoparticles relatively faint GFP expression only in photoreceptor cell layer was observed.
RGD peptide, transferrin and dual-functionalized anti-VEGF intraceptor plasmid-loaded nanoparticles systemic administration led to a statistically significant reduction in retinal (Figure 7a) and choroid-RPE (Figure 7b) VEGF levels 48 h after nanoparticle administration. Specifically, mean VEGF levels in the retina of the neovascular (right) eye of vehicle, naked plasmid or blank nanoparticle-treated Brown Norway (BN) rats were ~2.3-fold higher as compared to the control (left) eye. The neovascular eye retinal VEGF levels in the nonfunctionalized Flt23K-plasmid-loaded nanoparticle-treated group were approximately twofold higher as compared to the respective control eyes. On the contrary, the retinal VEGF levels in the neovascular eye of BN rats treated with RGD peptide, transferrin or dual-surface functionalized Flt23K-plasmid-loaded nanoparticles were comparable to those in the control eye.
Similar trend was observed for the VEGF levels in choroid-RPE of the neovascular eye. Specifically, mean VEGF levels in the choroid-RPE of the neovascular eye of vehicle, naked plasmid, blank nanoparticles and nonfunctionalized intraceptor plasmid-loaded nanoparticles were ~1.5-fold higher as compared to the control eye. The choroid-RPE VEGF levels in the neovascular eye of functionalized nanoparticle-treated groups were comparable to that of the control eye. Further, as expected, the control eye choroid-RPE VEGF levels were about fivefold higher as compared to the control eye retinal VEGF levels.
Anti-VEGF intraceptor plasmid-loaded surface-functionalized nanoparticles, but not the nonfunctionalized nanoparticles, led to significant reduction in the CNV area, 2 weeks after nanoparticle administration (Figure 8). Specifically, RGD peptide, transferrin and dual-functionalized nanoparticle-treated rats exhibited 73.3, 56.5 and 46.7% lower CNV areas in histopathology sections as compared to the ones treated with plasmid-loaded nonfunctionalized nanoparticles. As compared to the naked plasmid-treated group, RGD peptide, transferrin and dual-functionalized nanoparticle-treated groups had CNV areas reduced by 77.3, 62.9 and 54.5%, respectively.
Similar observation was made using choroidal flatmounts (Figure 9). Specifically, RGD peptide, transferrin and dual-functionalized nanoparticle-treated rats exhibited 54, 64 and 51% lower CNV areas in choroidal flatmounts as compared to the ones treated with plasmid-loaded nonfunctionalized nanoparticles. As compared to the naked plasmid-treated group, RGD peptide, transferrin and dual-functionalized nanoparticle-treated groups had CNV areas reduced by 64, 71 and 61%, respectively.
This article reports for the first time that (1) nanoparticles can be passively localized to CNV lesions on intravenous administration; (2) active targeting of nanoparticles by surface functionalization with a linear RGD peptide, transferrin, or both, significantly enhances the nanoparticle localization and subsequently gene delivery to CNV lesions on intravenous administration; (3) gene delivery to photoreceptors, RPE cells as well as endothelial cells can be accomplished using functionalized nonviral vectors and (4) anti-VEGF intraceptor plasmid loaded in functionalized nanoparticles is efficacious in inhibiting CNV. Thus, a systemically administered nonviral gene delivery system of anti-VEGF intraceptor plasmid is useful in the management of CNV, the cause of blindness due to wet AMD. Unlike the intravitreal or subretinal injections routinely used for gene delivery to the posterior segment of the eye, systemically administered delivery systems have the potential to increase safety and patient compliance.
In this study PLGA (50:50) nanoparticles were used to deliver anti-VEGF intraceptor plasmid intravenously. PLGA is a biodegradable polymer approved by FDA for human use. We have previously shown that PLGA nanoparticles cause approximately fourfold enhancement in the uptake of VEGF antisense oligonucleotide by RPE cells.37 Moreover, the inhibition of RPE cell VEGF secretion by oligonucleotide loaded in PLGA nanoparticles was observed to be similar to lipofectin-treated group. Cohen et al.38 have shown that PLGA nanoparticles maintain integrity of plasmid DNA and sustain the release of DNA for up to 28 days. Further, macromolecular delivery systems obtained by conjugating the therapeutic agent/dye to dextran have been reported to be delivered to CNV lesions due to enhanced permeation and retention effect.39–41 Therefore, we hypothesized that PLGA nanoparticulate system will be delivered to the CNV lesion upon intravenous administration.
In this study we used the widely accepted laser-induced CNV model. CNV was induced only in the right eye of the rats whereas the left eye served as control for each animal. To enable visualization of nanoparticles, all the nanoparticles were loaded with Nile red, in addition to the intraceptor plasmid. We consistently observed nanoparticles only in the neovascular eye flatmounts, but not in the control eye flatmounts (Figure 2). The delivery of nanoparticles to the neovascular eye can be attributed to the leaky blood–retina barrier as a result of CNV in the laser-treated eye. Laser photocoagulation has also been reported to enhance adenovirus-vector-mediated gene transfer in the rat retina on intravitreal injection.42 In this study, half of the posterior fundus (left hemisphere) was laser treated, whereas the other half was used as internal control. After intravitreal injection of adenoviral vector expressing β-galactosidase, the blue staining for LacZ was observed only in the laser-treated hemisphere but not in the other half. Indeed, in our studies we observed that the nanoparticles were distributed concentrically around the optic nerve (Figure 2a), exactly corresponding to the distribution of eight laser burns around the optic nerve. To identify all retinal cell types endocytosing nanoparticles, further experiments using multiple staining techniques need to be done in future.
On comparing the delivery of functionalized nanoparticles with that of nonfunctionalized ones, we observed that surface functionalization enhanced the nanoparticle delivery to the neovascular eye (Figure 3). The extent of enhancement and gene expression pattern was dependent on the functionalizing moiety (Figure 4). The nanoparticle delivery followed the order: Flt23K-NP (nonfunctionalized) <Tf-Flt23K-NP (transferrin functionalized) ~ RGD-Tf-Flt23K-NP (dual functionalized) <RGD-Flt23K-NP (RGD functionalized). The differences in the nanoparticle delivery as a function of targeting ligand might be due to the differences in the respective receptor expression. Another evidence to this effect is the gene expression pattern in various groups of nanoparticles. In case of RGD-functionalized nanoparticles in flatmounts, we observed very intense GFP expression in retinal vasculature and comparatively lesser expression in neural retina. However, with transferrin-functionalized nanoparticles we observed GFP expression mainly in neural retina. With dual-functionalized nanoparticles, a faint GFP expression was evident in retinal vasculature in addition to expression in neural retina. Indeed, VEGF has been reported to upregulate integrin receptors in retinal microvascular endothelial cells.43 VEGF expression is increased in laser-induced CNV.44 In addition, Yefimova et al.45 reported transferrin receptor immunoreactivity in multiple layers of neural retina, such as ganglion cell layer, inner nuclear layer, outer plexiform layer and photoreceptor inner segments.
Functionalization of nanoparticles with transferrin, RGD peptide, or both, does not enhance their nonspecific delivery to organs of reticuloendothelial system (Figure 5). In our biodistribution study we found that the delivery of functionalized nanoparticles to liver, lung, kidney, heart and spleen was less than or comparable to that of nonfunctionalized nanoparticles. The amount of nanoparticles detected in various tissues was in the nanogram range. This could be because of the tissue sampling time point of 24 h used in our studies. Indeed, Cheng et al.,46 as a result of their study of biodistribution of unconjugated and aptamer-conjugated PLGA nanoparticles as a function of time, found that the percentage of dose detected in various organs decreases with increasing time. Specifically, the percentage dose of PLGA nanoparticles detected in liver declined from ~20% at 2 h after intravenous injection to <0.1% at 24 h. This time-dependent decline in nanoparticle levels in liver can be explained by biliary excretion of nanoparticles subsequent to uptake by liver hepatocytes.47–49
Functionalized nanoparticles, but not nonfunctionalized nanoparticles, transfect RPE cells (Figure 6). To determine the retinal cell layers transfected by various treatments, DAPI-stained frozen sections of the enucleated eyes were examined by confocal microscopy for GFP expression 48 h after nanoparticle administration. Interestingly, ocular sections from rats treated with nonfunctionalized nanoparticles showed GFP expression only in the outer segments of photoreceptors, whereas sections from rats treated with transferrin, RGD peptide or dual-functionalized nanoparticles showed GFP expression in photoreceptors as well as RPE cell layer. GFP expression in RPE suggests that nanoparticles might have potentially been endocytosed in these cells. RPE cells do express and internalize integrin50 and transferrin receptors.51 As reported by Anderson et al.,50 αvβ3 and αvβ5 integrin receptors are expressed on the apical surface of RPE, whereas α5β1 integrin receptors are principally expressed on the basolateral surface of RPE. Further, integrin receptors are upregulated in ocular neovascularization. Specifically, integrin αvβ3 overexpression is observed in ocular tissues from AMD patients, whereas both integrins αvβ3 and αvβ5 are observed in vascular tissues from proliferative diabetic retinopathy patients.28 In primary cultures of human RPE cells, transferrin receptors have been reported to be expressed on both basolateral and apical surfaces.51 Moreover, both apical and basolateral transferrin receptors on RPE participate in receptor-mediated endocytosis through clathrin-coated pits, endosomes and tubular structures.51 Although the observation of GFP expression in photoreceptors as well as RPE warrants further studies, one possible explanation could be the retinal blood supply. Because the retinal blood supply consists of retinal vessels as well as choroidal vessels,52 nanoparticles can gain access into retinal tissues from both of these vasculatures after intravenous administration. Thus, the photoreceptor transfection might be occurring due to nanoparticles delivered by the retinal vessels whereas the RPE transfection might be due to the nanoparticles delivered from the choroidal side and subsequent endocytosis by RPE cells through cell-surface receptors. Indeed, Zhu et al.24 have reported that only immunoliposomes targeted with an antibody against transferrin receptors, but not the ones having a nonspecific immunoglobulin G on the surface, show expression of loaded β-galactosidase gene in the eye after intravenous administration. This suggests that presence of a functionalization moiety on the surface of nonviral vectors is critical in transfection in vivo.
Functionalized nanoparticles significantly reduced the retinal and choroid-RPE VEGF levels (Figure 7). To assess the activity of anti-VEGF intraceptor in vivo, VEGF levels in the retina and choroid-RPE of BN rat eyes were quantified by a sandwich enzyme-linked immunosorbent assay (ELISA) method 48 h after nanoparticle administration. As expected, retinal as well as choroid-RPE VEGF levels in the laser-treated eye were significantly higher as compared to the control eye in vehicle-treated rats.44,53 Similarly, VEGF levels were high in groups treated with naked plasmid, blank nanoparticles or nonfunctionalized Flt23K-plasmid-loaded nanoparticles. On the other hand, rats treated with transferrin, RGD peptide or dual-functionalized Flt23K-plasmid-loaded nanoparticles exhibited lower retinal and choroid-RPE VEGF levels in the laser-treated eye that were comparable to the control eye. Because laser-treated retinal VEGF levels are reduced to normal but not lower levels on intraceptor treatment, intraceptor expression in photoreceptors (Figure 6) is unlikely to perturb the physiology of the normal retina. A significant outcome of these observations is that we have a nanoparticle system that can transfect photoreceptor cells in vivo, which is not feasible for several other nonviral gene delivery systems. 54 The choroid-RPE VEGF levels we observed were approximately fivefold higher compared to the retinal VEGF levels. This observation is in agreement with the literature on polarized VEGF secretion by RPE cells, wherein 2- to 7-fold higher VEGF secretion toward the basolateral side has been reported.55 Further, our retinal VEGF levels in BN rat control eyes are in good agreement with the published literature.56
Most significantly, functionalized nanoparticles were found to be efficacious in reducing laser-induced CNV areas as compared to vehicle- and nonfunctionalized-nanoparticle-treated groups (Figures 8 and and9).9). CNV areas were quantified 2 weeks after treatment administration in histopathology sections and choroidal flatmounts. This is the first report documenting efficacy of anti-VEGF intraceptor plasmid in inhibiting CNV. Further, targeted nanoparticles can be used as delivery systems to deliver this plasmid to the CNV lesions.
Because none of the treatment approaches is free of side effects, before any clinical assessment, we will need to study the safety of intravenous administration of plasmid-loaded nanoparticles for gene delivery to the eye. Although the possibility of inflammatory response due to polymeric particles cannot be ruled out, the extent of such a reaction depends on the nature of the polymer. Results of comparison of proinflammatory potential of nonbiodegradable and biodegradable PLGA nanoparticles were recently published.57 In general, a significantly lower immune response was observed with biodegradable PLGA nanoparticles in lungs. The difference in the inflammatory response of biodegradable nanoparticles as compared to that of nonbiodegradable nanoparticles was most prominent in terms polymorphonucleocyte recruitment. In case of animals injected with polystyrene nonbiodegradable nanoparticles, up to 74% polymorphonucleocytes could be detected in total bronchoalveolar lavage cell population. On the other hand, in case of animals injected with PLGA nanoparticles, polymorphonucleocyte recruitment did not exceed that of isotonic glucose (negative control) group. Further, we have previously assessed the safety of periocularly administered PLGA microparticles in Sprague–Dawley rats.58 In hematoxylin-and-eosin (H&E)-stained histological sections of the eye, no cellular infiltration, inflammation or atrophy of retina was observed. Thus, it can be concluded that PLGA nanoparticles do not likely lead to a serious inflammatory response.
In conclusion, the results documented herein open up an invaluable avenue in the management of wet AMD by providing a novel therapeutic gene, a safer route of administration and also a nonviral delivery system for gene delivery. Although the safety of this system needs to be studied in detail, the tremendous potential of functionalized nanoparticles in gene delivery to the eye by intravenous administration cannot be overlooked. The clinical application of this system goes beyond delivery of anti-VEGF intraceptor plasmid and includes the application of the functionalized nanoparticles for the delivery of drugs/macromolecules to inhibit CNV in a targeted manner. Systemic nonviral gene therapy with long-term effect may offer a more convenient, less risky and less expensive modality for the treatment of AMD.
Anti-VEGF intraceptor plasmid was kindly provided by Dr Balamurali Ambati (Moran Eye Institute, Salt Lake City, UT, USA). Endofree Plasmid Giga kit was obtained from Qiagen Inc. (Valencia, CA, USA). PLGA 50:50 (Resomer RG 503H) was obtained from Boehringer Ingelheim, Germany. RGD peptide, apo-transferrin, polyvinyl alcohol (PVA), ketamine/xylazine, 3-(N-Morpholino) propanesulfonic acid (MOPS) and N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDAC) were obtained from Sigma-Aldrich (St Louis, MO, USA). Methylene dichloride, high-pressure liquid chromatography grade, was purchased from Fisher Scientific Inc. (Pittsburgh, PA, USA).
Anti-VEGF intraceptor plasmid was amplified using Endofree Plasmid Giga kit (Qiagen Inc.). Plasmid-loaded Nile red PLGA nanoparticles were prepared by double emulsion solvent evaporation method, as previously described with slight modifications. 37 Briefly, PLGA equivalent to 25 times the plasmid amount was dissolved in 1 ml methylene dichloride. To enable confocal fluorescent visualization of nanoparticles, Nile red, 250 µg, was added to PLGA solution. Plasmid solution was added to the organic phase (aqueous/organic = 1:3), and emulsified by probe sonication at 6W on ice for 1 min. This w/o emulsion was poured into six times volume of 2% w/v PVA, and probe was sonicated at 24 W for 3 min to form a w/o/w double emulsion. The emulsion was stirred in excess of 2% w/v PVA solution at room temperature for 3 h. The nanoparticle suspension was subsequently subjected to rotary evaporation (40 °C; 2 h). Nanoparticles were retrieved by ultracentrifugation (34 155 g; 0.5 h). The particles were washed with deionized water thrice by repeated dispersion and ultracentrifugation cycles. Finally, the particles were lyophilized.
Nanoparticles were conjugated to RGD peptide (Gly-Arg-Gly-Asp-Ser-Pro-Lys) and apo-transferrin using carbodiimide chemistry as previously reported22 with slight modifications. Briefly, nanoparticles were suspended in MOPS buffer (50 mm) pH 6.5. The surface carboxylate functional groups were activated by incubation with EDAC (0.01 m) at room temperature for 2 h with vortexing. Ligand solution (10 µm) in MOPS buffer (pH 6.5) was added dropwise to activated nanoparticles and vortexed at room temperature for 12 h. Conjugated nanoparticles were separated by ultracentrifugation (34 155 g; 0.5 h) and were lyophilized.
Nanoparticles (1:1000 dilution in filtered deionized water) were characterized for their size and ζ-potential using dynamic light scattering (ZetaPlus zeta potential analyzer; Brookhaven Instruments Ltd, New York, NY, USA). For plasmid loading determination, a known amount of lyophilized nanoparticles was dissolved in dichloromethane. The plasmid was then extracted into twice the volume of TE buffer. The absorbance of the aqueous layer at 260 nm was measured and used to calculate the amount of plasmid per mg of nanoparticles. To determine the propensity of nanoparticles to aggregate, the nanoparticles (10 mg) were incubated with either PBS (pH 7.4) or serum for 48 h at 37 °C. At 0, 24 and 48 h particle size was determined using dynamic light scattering.
All animals were treated according to the Association for Research in Vision and Ophthalmology (ARVO). Statement for the Use of Animals in Ophthalmic and Vision Research. The protocols for this study were approved by the Institutional Animal Care and Use Committee of the University of Nebraska Medical Center. CNV was induced as described previously.59 Adult male BN rats (150–180 g) were purchased from Harlan Sprague Dawley Inc. (Indianapolis, IN, USA). Rats were anesthetized using an intraperitoneal injection of 40–80 mg ml−1 ketamine and 10–12 mg kg−1 xylazine mixture. Pupils were dilated by topical administration of 1% tropicamide. The fundus was visualized using a coverglass and 2.5% hypromellose solution (Gonak; Akorn Inc., Buffalo Grove, IL, USA). Eight laser spots (100 µm, 150 mW, 100 ms) concentric with the optic nerve were placed in the right eye of each rat using a 532 nm diode laser (Oculight Glx; Iridex Inc., Mountain View, CA, USA) and a slit lamp (Zeiss slit lamp 30SL; Carl Zeiss Meditec Inc., Dublin, CA, USA). Left eye was used as a control for each animal. The Bruch’s membrane breakage was confirmed by the end point ‘bubble formation’. Rats showing intraocular hemorrhage on laser administration were excluded from the study.
On day 14 after laser treatment, the rats were administered nanoparticles (10 mg dose; 250 µl injection volume) by injection into the tail vein. Eyes were enucleated 24 (N = 3) and 48 h (N = 3) after nanoparticle administration. The 24 h eye samples were used to assess nanoparticle delivery in posterior segment flatmounts, whereas 48 h eye samples were subjected to frozen sectioning.
For preparation of flatmounts, eyes were fixed in 4% paraformaldehyde for 1 h immediately after enucleation. Anterior segment, lens and vitreous were removed. The remaining eyecup was washed with ICC buffer (0.5% bovine serum albumin, 0.2% Tween 20, 0.05% sodium azide in PBS).60 The eyecups were subsequently stained for cell nuclei by incubating with 1:1000 dilution of 5 mg ml−1 solution of DAPI (Invitrogen-Molecular Probes, Eugene, OR, USA) in ICC buffer for 4 h at 4 °C with gentle rotation. Eyecups were again washed with ICC buffer. Radial cuts were made toward the optic nerve head, sclera–choroid–retina complexes were flatmounted using Gel-mount (Biomeda Corp., Foster City, CA, USA).
Frozen sectioning of eyes was performed as reported previously.61 Briefly, enucleated eyes were fixed for 2 h in 4% paraformaldehyde and rinsed with PBS. The eyes were subsequently cryoprotected by incubation overnight in 30% sucrose. Eyes were then embedded in optimal cutting temperature compound medium (Sakura Finetek Inc., San Diego, CA, USA). Serial sections (10 µm thick) were obtained using a cryostat and then thaw-mounted onto gelatin-coated slides. The slides were stored at −20 °C until they were processed for confocal microscopy. Sections were hydrated for 5 min in PBS and stained with DAPI using Vectashield mounting media (Vector Laboratories Inc., Burlingame, CA, USA).
Confocal imaging was performed with a Zeiss LSM 510 Meta confocal microscope equipped with Zeiss LSM software for image reconstruction (Carl Zeiss, Jena, Germany). All sections were scanned in multitrack mode to avoid overlap of red (excitation 561 nm), green (excitation 488 nm) and blue (excitation 405 nm) channels. Flatmounts were imaged using × 20 objective, whereas frozen sections were imaged using × 40 objective.
The biodistribution of nonfunctionalized and functionalized nanoparticles was determined by measuring the Nile red fluorescence in various tissues 24 h after injection of Nile-red-loaded nanoparticles into the tail vein of BN rats. Specifically, day-14 laser-treated BN rats were divided into four groups (N = 6 per group) to receive one of the following types of nanoparticles: (1) Nile-red-loaded PLGA nanoparticles (NP); (2) transferrin functionalized Nile-red-loaded PLGA nanoparticles (Tf-NP); (3) RGD peptide functionalized Nile-red-loaded PLGA nanoparticles (RGD-NP) and (4) RGD peptide-and transferrin-functionalized Nile-red-loaded PLGA nanoparticles (RGD-Tf-NP). Nile red loading in the nanoparticles was 890 µg mg−1 of nanoparticles. A nanoparticle dose of 10 mg in 250 µl of PBS (pH 7.4) was injected into the tail vein of the rats. The rats were euthanized at 24 h after nanoparticle administration. Various tissues, such as liver, spleen, kidney, lung, heart and brain, were isolated. Nile red was quantified by extracting the dye into dichloromethane (5 ml) after homogenizing the tissue in PBS (1 ml). The organic layer was pipetted out and dichloromethane was evaporated under a stream of nitrogen. The residue was reconstituted in 1 ml of dichloromethane. The fluorescence of Nile red was measured at an excitation and an emission wavelength of 544 and 590 nm, respectively.
On day 14 after laser induction, the rats were administered one of the following treatments: (1) vehicle, (2) naked Flt23K plasmid, (3) blank nanoparticles, (4) nonfunctionalized Flt23K-plasmid-loaded nanoparticles (Flt23K-NP), (5) transferrin-conjugated Flt23K-plasmid-loaded nanoparticles (Tf-Flt23K-NP), (6) RGD-peptide-conjugated Flt23K-plasmid-loaded nanoparticles (RGD-Flt23K-NP), (7) dual (RGD peptide and transferrin)-conjugated Flt23K-plasmid-loaded nanoparticles (RGD-Tf-Flt23KNP). Eyes were enucleated 2 days and 2 weeks after nanoparticle administration. The 2-day eye samples were used to assess the effect of various treatments on retinal and choroid-RPE VEGF levels, whereas the 2-week samples were used for histopathologic examination and choroidal flatmounts.
For VEGF quantification, the eyes were enucleated 2 days after nanoparticle administration. The retina and choroid-RPE were isolated. The tissues were homogenized in PBS (pH 7.4). The homogenate was centrifuged at 10 000 r.p.m. for 5 min. VEGF levels in the clear supernatant thus obtained were quantified using sandwich ELISA method as described earlier.58
The histopathologic examination of CNV lesions was performed 2 weeks after nanoparticle intravenous administration. The eyes were enucleated after killing the animals and fixed in 4% paraformaldehyde overnight. Paraffin sections (6 µm thick) stained with H&E were imaged using a light microscope. The images were examined by a masked ophthalmic pathologist (HEG). To enable quantitative comparison across various groups, the CNV area was quantified using ImageJ software (NIH).
For choroidal flatmount, the rats after killing were infused with 10 ml of PBS (pH 7.4). This was followed by infusion with 10 ml 4% paraformaldehyde. Finally, 4 ml of 50 mg ml−1 fluorescein isothiocyanate (FITC)-dextran solution (2 × 106 Da) was infused. The eyes were then enucleated. Choroidal flatmounts were prepared and analyzed by a masked ophthalmic pathologist (HEG). The flatmounts were imaged with a Nikon EZ-C1 confocal microscope using 488 and 568 nm excitation wavelengths at ×200. CNV areas were obtained using ImageJ software.
Comparison of means of various groups was done using nonparametric statistical analysis. Comparison of two groups was carried out using Mann–Whitney test, however for comparison of more than two groups, Kruskal–Wallis nonparametric analysis of variance was used. Differences were considered statistically significant at P<0.05.
This work was primarily supported by NIH grants R24 EY017045 and R21 EY017360. Creation of VEGF intraceptor plasmid by Dr BK Ambati was supported by NIH grant 5RO1EY017182. We thank James R Talaska and Janice Taylor of the Confocal Laser Scanning Microscopy Core Facility at University of Nebraska Medical Center, which is supported by the Nebraska Research Initiative, for providing assistance with confocal microscopy. We thank Karen Dulany and Maureen Harman of the Eppley Histology Core Laboratory at University of Nebraska Medical Center, for their help in cryosectioning of the tissues. We also thank Dr Chandrasekar Durairaj and Rajendra S Kadam for their assistance during the study. We especially thank Dr Weiqing Gao, Emory Eye Center, Emory university, Atlanta, GA, for assistance with choroidal flatmounts.