Journal of Controlled Release 200 (2015) 115–124

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Transcorneal iontophoresis of dendrimers: PAMAM corneal penetration and dexamethasone delivery Joel G. Souza a, Karina Dias a, Silas A.M. Silva b, Lucas C.D. de Rezende a, Eduardo M. Rocha c, Flavio S. Emery a, Renata F.V. Lopez a,⁎ a b c

Department of Pharmaceutical Sciences, School of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Ribeirão Preto, SP, Brazil Medicine Department, Federal University of São Paulo, SP, Brazil Department of Ophthalmology, Otorhinolaryngology and Head & Neck Surgery, School of Medicine at Ribeirão Preto, University of São Paulo, Ribeirão Preto, SP, Brazil

a r t i c l e

i n f o

Article history: Received 22 October 2014 Accepted 24 December 2014 Available online 29 December 2014 Keywords: Transcorneal iontophoresis Dendrimer Dexamethasone sustained delivery Confocal microscopy Nanoparticle characterization

a b s t r a c t Iontophoresis of nanocarriers in the eye has been proposed to sustain drug delivery and maintain therapeutic concentrations. Fourth generation polyamidoamine (PAMAM) dendrimers are semi-rigid nanoparticles with surface groups that are easily modified. These dendrimers are known to modulate tight junctions, increase paracellular transport of small molecules and be translocated across epithelial barriers, exhibiting high uptake by different cell lines. The first aim of this study was to investigate the effect of iontophoresis on PAMAM penetration and distribution into the cornea. The second aim was to evaluate, ex vivo and in vivo, the effect of these dendrimers in dexamethasone (Dex) transcorneal iontophoresis. Anionic (PAMAM G3.5) and cationic (PAMAM G4) dendrimers were labeled with fluorescein isothiocyanate (FITC), and their distribution in the cornea was investigated using confocal microscopy after ex vivo anodal and cathodal iontophoresis for various application times. The particle size distribution and zeta potential of the dendrimers in an isosmotic solution were determined using dynamic light scattering and Nanoparticle Tracking Analysis (NTA), where the movement of small particles and the formation of large aggregates, from 5 to 100 nm, could be observed. Transcorneal iontophoresis increased the intensity and depth of PAMAM-FITC fluorescence in the cornea, suggesting improved transport of the dendrimers across the epithelium toward the stroma. PAMAM complexes with Dex were characterized by 13C-NMR, 1H-NMR and DOSY. PAMAM G3.5 and PAMAM G4 increased the aqueous solubility of Dex by 10.3 and 3.9-fold, respectively; however, the particle size distribution and zeta potential remained unchanged. PAMAM G3.5 decreased the Dex diffusion coefficient 48-fold compared with PAMAM G4. The ex vivo studies showed that iontophoresis increased the amount of Dex that penetrated into the cornea by 2.9, 5.6 and 3.0fold for Dex, Dex-PAMAM G4 and Dex-PAMAM G3.5, respectively. In vivo experiments, however, revealed that iontophoresis of Dex-PAMAM-G3.5 increased Dex concentration in the aqueous humor by 6.6-fold, while iontophoresis of Dex-PAMAM G4 and Dex increased it 2.5 and 2-fold, respectively. Therefore, iontophoresis targeted PAMAM to the cornea but it is the sustained delivery of the Dex from PAMAM that prevents its rapid elimination from the aqueous humor. In conclusion, iontophoresis of PAMAM complexes represents a promising strategy for targeted and sustained topical drug delivery to the eye. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Transcorneal and transcleral iontophoresis are non-invasive and safe physical methods that involve the application of a low electrical potential gradient across the cornea [1–3] and sclera [1,4,5], respectively, to enhance drug delivery to anterior and posterior segments of the eye. Early studies have demonstrated that iontophoresis increased the concentration of penicillin in the aqueous humor compared with subconjunctival injections [6] and recently, successful delivery of nucleic acids [7,8] and ⁎ Corresponding author at: School of Pharmaceutical Sciences of Ribeirão Preto, University of São Paulo, Av. do Café, s/n, 14040-903 Ribeirão Preto, SP, Brazil. E-mail address: [email protected] (R.F.V. Lopez).

http://dx.doi.org/10.1016/j.jconrel.2014.12.037 0168-3659/© 2015 Elsevier B.V. All rights reserved.

siRNAs [9] into the eye has been reported. In general, iontophoresis contributes to drug delivery mainly by electromigration, which is the orderly movement of ions in the presence of an electrical current, and electroosmosis, which is convective solvent flow in the anode-to-cathode direction that occurs under physiological conditions when the epithelium is negatively charged [10]. The contribution of electromigration in the ocular iontophoresis of ionized drugs has been explored [11–15]. The effect of electroosmosis, however, which is currently considered to be the dominant mechanism for uncharged and high-molecular-weight drugs [4,9], remains underexplored. Although iontophoresis is able to increase the drug concentration in the eyes, a limitation of its use is the rapid clearance of the drug from eye tissues, thus requiring repeated iontophoretic applications [14].

116

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

Therefore, nanocarriers have been evaluated with iontophoresis to provide sustained drug delivery in different regions of the eye [16]. Currently, iontophoretic transport of two different nanocarriers has been studied: 1) cationic- and anionic-fluorescent-modified polystyrene nanospheres (FluoSpheres®) of approximately 20 nm (anionic) and 40 nm (cationic) were studied to explore the contribution of electromigration to their transcorneal and transcleral delivery [16] and 2) anionic micelles of approximately 4.5 nm in size were studied to explore their potential to prolong dexamethasone (Dex) delivery using cathodal transcleral iontophoresis [14]. In this work, we explore the potential of polyamidoamine (PAMAM) dendrimers in association with iontophoresis for delivery of Dex into/ through the cornea. PAMAM dendrimers are well defined and highly repetitively branched synthetic globular structures. These structures can be readily modified by, for example, increasing the dendrimer generation, or by changing their surface groups [17]. These modifications can change the ability of the dendrimers to interact with biological membranes or to be taken up by cells [18,19]. These modifications can also improve drug encapsulation or interactions, thus changing the drug's solubility, stability, permeability and distribution throughout the site of administration [20–23]. Specifically related to the evaluation of dendrimers as topical ocular drug carriers, it has been demonstrated that PAMAM generations 2, 3 and 4 (G2, G3 and G4) increased the bioavailability of pilocarpine nitrate and tropicamide relative to the administration of these drugs in a phosphate buffer solution [17]. In addition, PAMAM G4 and G5 with amine or carboxyl groups increased the half-life of puerarin in rabbit eyes [24]; and the pharmacokinetic properties of puerarin were improved in aqueous humor when it was complexed with PAMAM G3 [25]. Despite these satisfactory results, the effect of iontophoresis on PAMAM ocular penetration has never been studied. We hypothesize that if iontophoresis can facilitate the transport of dendrimers into the cornea, then these nanostructures could provide a means for localized and sustained drug release. The drug selected for our study was Dex because it is widely used for the treatment of uveitis and other inflammatory eye conditions. However, its high lipophilicity prevents the development of concentrated aqueous formulations, and protective mechanisms of the eyes drain it quickly from the administration site, thus requiring frequent administrations. Our working hypothesis, therefore, is that Dex-PAMAM complexes will sustain Dex delivery and increase Dex solubility [24], enabling the administration of higher doses with less frequent administration. Moreover, iontophoresis can facilitate transport of the complexes to the cornea providing a targeted and sustained release of Dex. Therefore, the first aim of this work was to verify, for the first time, the effect of iontophoresis on PAMAM cornea penetration by using PAMAM labeled with FITC. PAMAMs with 64 surface groups (generation 4) and different surface charges (cationic and anionic) were used. The second aim was to prepare and characterize Dex-PAMAM complexes and to evaluate the effect of iontophoresis on the amount of Dex that permeates the cornea using ex vivo and in vivo models.

2. Materials and methods 2.1. Material Cationic (PAMAM G4, linear formula: [NH2(CH2)2NH2]:(G = 4);dendri PAMAM(NH2)64) and anionic (PAMAM G3.5, linear formula: [NH2(CH2)2NH2]:(G = 3.5);dendri PAMAM(NHCH2CH2COONa)64) generation 4 PAMAM dendrimers were purchased from Sigma Chemical Co (St Louis, MO). PAMAM G4 has 64 primary amines as surface groups, while PAMAM G3.5 has 64 carboxyl groups as surface groups. Their size, distribution and zeta potentials when dispersed at 1 mg/mL in isotonic PBS, 50 mM, pH 7.4, were determined using a Zetasizer Nano ZS 90 (Malvern Instruments, UK) at 25 °C and a Nanosight (LM 20, UK).

Dex base was purchased from Tianjin Pharmaceutic (China), sodium mono and dihydrogen phosphate and sodium chloride were purchased from Synth (Diadema, Brazil); acetonitrile (ACN) and N2-hydroxyethylpiperazine-N-2-ethanesulphonic acid (HEPES) were purchased from J.T. Baker (Phillipsburg, NJ, USA); and fluorescein isothiocyanate (FITC), N-Ethyl-N′-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and ethylenediamine (EDA) were purchased from Sigma Chemical Co (St Louis, MO). Deionized water (18.2 MΩ-cm at 25 °C) (Milli-Q, Direct-Q 3 UV, Millipore, Billerica, MA, USA) was used to prepare all of the solutions. 2.2. FITC-labeled PAMAM dendrimers FITC-labeled PAMAM G4 and G3.5 dendrimers were prepared according to the literature [19,20,26,27]. To evaluate the number of FITC molecules that reacted with PAMAM terminal groups, the fluorescence intensity of FITC was monitored using a spectrofluorimeter (F4500, HITACHI, Tokyo, Japan) at 490/518 nm (excitation/emission). For both PAMAMs, a media of 1.2 FITC molecules covalently bonded to dendrimers was observed. The FITC-labeled PAMAMs were freezedried to obtain the conjugates in powder form. The absence of nonreacted FITC was confirmed by HPLC. A C18 column (Shim-Pack VPODS, 4.6 mm × 25 cm, Shimadzu, Tokyo, Japan) was employed for this purpose. The mobile phase was composed of acetic acid (1% v/v) in water and ACN (10:90), was used at a flow rate of 1 mL/min. The excitation/emission wavelengths were 495/525 nm. Fluorescence was detected using a fluorescence detector (RF-10AXL, SHIMADZU, Shimadzu, Tokyo, Japan). Under these conditions, the non-reacted FITC has a retention time of 4 min. Therefore, any non-reacted FITC could be detected by the presence of a peak at 4 min. 2.3. PAMAM ex vivo distribution in the cornea The corneas used in the experiments were obtained from pig eyes that had been collected immediately after the slaughter of the animal (Frigorífico Pontal Ltda, Pontal, Brazil). After removal from the animals, the eyes were maintained at 4 °C while being transported to the laboratory and were used within 1.5 h of enucleation. Corneoscleral buttons were dissected using standard eye bank techniques, and care was taken to minimize tissue distortion [28]. After the preparation of the cornea, modified iontophoretic vertical diffusion cells [3] were assembled with the cornea, separating the donor from the receptor compartment. The area of the cornea that was exposed was 0.78 cm2. The receptor was filled with isotonic HEPES buffer (25 mM, pH 7.4) and maintained at 35 °C. The donor compartment was filled with 1 mL of PAMAM G3.5-FITC or G4-FITC dispersed in isotonic PBS, pH 7.4, at 200 μg/mL of FITC. A control experiment was performed with a saturated PBS solution of free FITC. A positive electrode (Ag) was placed in contact with the PAMAM G4FITC dispersion in the donor compartment (anodal iontophoresis), and the negative counter-electrode (AgCl) was placed in contact with the receptor solution to complete the circuit. Corneal penetration of the negatively charged PAMAM G3.5-FITC and FITC was investigated by cathodal and anodal iontophoresis to evaluate the contribution of electromigration and electroosmosis, respectively. A constant electrical current of 1 mA/cm2, generated by a power supply (Kepco APH 500 DM, Kepco, USA), was applied for 5 min or 3 h to evaluate the effect of iontophoresis application time on PAMAM corneal penetration. The same protocol, but in the absence of the electrical current, was performed for the passive experiments. The integrity of the cornea following iontophoresis was previously confirmed after 6 h of electrical current application, with a total electric current density of 180 mA/cm2 [3]. Fluorescence of the corneal samples after the ex vivo passive and iontophoretic experiments was preserved using Tissue Tek solution (O.C.T. compound). The corneas were frozen at −80 °C. For the analyses of fluorescence distribution in the cornea, two different protocols were

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

used. First, prior to microscopic analysis 20 μm cryosections that were obtained perpendicular to the corneal surface were examined employing a Leica TCS SP5 confocal microscope (Mannheim, Germany) with a 20× immersion objective. The samples were excited with a laser at 480 nm, and the fluorescence was detected at 515–530 nm. In the second protocol, the corneas were inserted into an image chamber (Coverwell™ imaging chamber, Invitrogen), and the confocal images were taken in a plane parallel to the corneal surface (xymode) with a 40× immersion objective using LAS AF Lite 2.6.0 software (Leica, Germany). Optical sections were prepared in the z-stack mode, fixing the corneal surface as the initial position. The samples were excited as described above. Corneal samples without treatment were used as negative controls to configure the microscope settings. 2.4. Dex-PAMAM complexes The solid complexes of Dex-PAMAM were prepared according to the literature [20,24]. First, 116 mg of Dex (MW = 392.46) was dissolved in 25 mL of methanol. Next, 42 mg of PAMAM G4 (MW = 14214.17) or 38 mg of PAMAM G3.5 (MW = 12927.69) was added to obtain a molar ratio of 100 mol of Dex for each mole of dendrimer. The reaction mixture was stirred at 300 rpm for 24 h, while being protected from the light, followed by evaporation of the methanol at room temperature. Next, water was added and the dispersion was stirred at 300 rpm for 24 h. The suspension formed was centrifuged for 10 min at 3000 rpm, and the supernatant was filtered using 0.45-μm PVDF membranes. The filtered solution containing the Dex-PAMAM complex was frozen at −80 °C and freeze-dried (K105, Liotop, São Carlos, SP, Brazil) to obtain the complex in powder form [22,24]. To determine the Dex/PAMAM molar ratio after complex formation, the complexes were weighed and dissolved in water. The samples were diluted accordingly and analyzed using a UV/Vis spectrophotometer at 260 nm because in this region, the dendrimers exhibit no significant absorption. The effect of dendrimers on the water solubility of Dex was evaluated by analyzing the absorbance of a Dex-PAMAM complex solution and comparing the results with the absorbance of a filtered saturated Dex aqueous solution. 2.5. Characterization of complexes The formation of the complexes was confirmed by 1H-NMR, 13CNMR and DOSY. The analyses were performed using a Bruker DRX 500-MHz NMR system from Bruker Daltonics® (Billerica, MA, USA). The size, distribution and zeta potential of the dendrimers free, or complexed with Dex (dispersed in isotonic PBS, pH 7.4, at 1 mg/mL) were analyzed by dynamic light scattering (DLS) using a Zetasizer Nano ZS 90 equipped with a He–Ne laser with a wavelength of 633 nm, a detector angle of 90° and a sample volume of 1 mL (Malvern Instruments, UK) at 25 °C. In addition to DLS, the particle size was determined with Nanosight (LM 20, UK) using the Nanoparticle Tracking Analysis software (NTA 2.0). The osmolality of isotonic PBS, pH 7.4, saturated with Dex, DexPAMAM G4 or Dex-PAMAM G3.5 was determined using an osmometer (Semi-micro osmometer, model K 7400, Knauer) by measuring the decrease in the melting point of each dispersion.

117

a HEPES buffer solution, pH 7.4. This solution was maintained under magnetic stirring (600 rpm) at 35 °C. Samples of 1 mL were collected from the receiving solution for 12 h at 1 h intervals for the first 6 h, then at 2 h intervals after 6 h. After each withdrawal, the receptor was replaced with fresh receiving fluid. The amount of Dex in the receptor solution was determined by LC-MS, as described in Section 2.9. The Dex diffusion coefficients (D) in Dex solutions or in complex dispersions were calculated using Eq. (1) [29]: 1=2

Q ¼ 2Co ðDt=πÞ

;

ð1Þ

where Q is the cumulative amount of drug released per unit area, Co is the initial drug concentration in the donor compartment, D is the drug diffusion coefficient and t is the time. The drug release rate (K) was calculated from the slope of plots of Dex permeated versus the square root of time. 2.7. Dex ex vivo corneal iontophoresis Experiments were performed as described in Section 2.3, except that the donor compartment (anode or cathode) was filled with Dex or DexPAMAM complexes. Thus, 1 mL of isotonic PBS pH 7.4 saturated with Dex-PAMAM G3.5 or Dex-PAMAM G4 (containing approximately 940 μg/mL and 353 μg/mL of Dex, respectively) was placed in contact with the cornea. A suspension containing 1 mg/mL Dex was also evaluated. The anodal transport of Dex was analyzed as a function of time for all formulations. Dex permeation from Dex-PAMAM G3.5 was also analyzed after cathodal iontophoresis. A constant electrical current of 1.0 mA/cm2 generated by a Kepco APH 500DM apparatus (Kepco Power Supply, USA) was applied. Samples of 1 mL were withdrawn from the receiving solution each hour for 3 h and replaced by fresh receiving fluid. “Passive” experiments were performed under conditions identical to those described above, except that no current was applied. The cumulative amount of Dex that permeated the cornea was plotted versus time. For quantification, Dex was recovered from the receptor solution by partitioning with ethyl acetate by mixing 1 mL of the receptor solution with 1 mL of ethyl acetate in a vortex. After partitioning, 700 μL of the ethyl acetate phase was collected, evaporated and re-suspended in 700 μL of methanol:water (50:50) for LC-MS analysis (Section 2.9). After the 3 h ex vivo experiments, Dex was recovered from the cornea by cutting it into small pieces and transferring the pieces to a 50-mL Falcon tube with 5 mL of methanol:water (50:50). The cornea was then homogenized using a tissue homogenizer (Turratec TE-102; Tecnal, Brazil) for 1 min at 17,500 rpm and then sonicated in an ultrasonic bath for 5 min. The samples were centrifuged at 13,000 ×g in an ultracentrifuge (Thermo, Megafuge 16R), filtered through 0.45-μm PVDF membranes and analyzed by LC-MS (Section 2.9). The efficiency of Dex recovery from the cornea and receptor solutions was calculated by loading either pieces of cornea, or isotonic PBS, with known quantities of Dex in methanol. After solvent evaporation, the extraction process described above was performed, and the amount of Dex recovered from each matrix was determined. The Dex recoveries from the cornea and PBS were within the range of 95.7– 99.1% and 101.42–101.57%, respectively (n = 3 in both cases). The coefficients of variation were below 5%.

2.6. In vitro release studies of Dex from Dex-PAMAM complexes 2.8. Dex in vivo corneal iontophoresis Release rates of Dex from the complexes or from Dex saturated solutions were determined using a vertical diffusion cell [28] and a cellulose acetate membrane (cutoff = 3500 Da) as a support for the formulations. The PAMAM complexes or Dex solutions were maintained under conditions of Dex saturation, resulting in Dex concentrations of 91.7, 353.1 and 939.8 μg/mL for Dex, Dex-PAMAM G4 and Dex-PAMAM G3.5, respectively. Next, 1 mL of the formulations was placed in the donor compartment, and the receptor compartment was filled with 35 mL of

2.8.1. Animals New Zealand rabbits weighing 2.0–2.5 kg were used as the experimental model. The protocol was approved by the Ethics Committee of the University of São Paulo (Protocol no. 11.1.121.53.6), which is in accordance with the National Institutes of Health (NIH) Guidelines for the Care and Use of Laboratory Animals. During the experiments, rabbits were anesthetized by intramuscular application of a mixture containing

118

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

ketamine hydrochloride (35 mg/kg) and xylazine (3.5 mg/kg) every 40 min. At the end of the experiments, animals were sacrificed in a saturated CO2 chamber.

3. Results and discussion

2.8.2. In vivo iontophoresis application After anesthetizing the animals, a silicone template was carefully placed in contact with the animals' cornea. The area of the cornea outlined by the device was approximately 0.8 cm2. The template border that was in immediate contact with the cornea was coated with silicone grease to prevent formulation leakage. Afterwards, 1 mL of the saturated dispersions of Dex free or complexed with PAMAM G4 was added to the template and exposed to the positive electrode. A saturated DexPAMAM G3.5 dispersion was placed in contact with the negative electrode. A constant electric current of 5.1 mA/cm2 was applied using a Kepco APH 500DM apparatus (Kepco Power Supply, USA) for 4 min. Note that the same electric current density was used by other authors [30] for transcorneal delivery of Dex phosphate. To complete the circuit, a patch (Iomed, Inc., Salt Lake City, Utah, USA) was fixed to the animal's ear as the counter electrode. After application of iontophoresis, the animals were maintained anesthetized for 2 h to permit posterior collection of the aqueous humor and quantification of Dex in the samples. Passive experiments were performed in the same manner, but without the application of an electric current. Four animals were used in each experiment. At the end of the experiments, animals were sacrificed in a CO2 chamber and the aqueous humor was processed for Dex quantification. A 100 μL sample was transferred to a 1.5 mL Eppendorf tube and partitioned using 200 μL of ethyl acetate. The tubes were vortexed for 1 min then allowed to stand undisturbed. After partitioning, 150 μL of the ethyl acetate phase containing Dex was collected and the solvent was evaporated. The samples were dissolved in 100 μL of water: MeOH (1:1) and analyzed by LC-MS.

The sizes of the PAMAM nanoparticles were first determined by DLS. Currently, DLS is the most frequently used technique for determining nanoparticle sizes [32], and it has been employed for measuring the size of PAMAM dendrimers [33,34]. However, in our study, the size distribution of PAMAM G4 and PAMAM G3.5 in PBS was observed to be polydispersed and multimodal. Particles ranging from 4 nm to more than 250 nm could be identified in the intensity distribution spectra. The percentage of different size nanoparticles also changed with time. Specifically, during the analysis, at some times nanoparticles of approximately 4 nm could be seen in 80% of the population, and at other times 250 nm populations were predominant. According to the literature, generation 4 PAMAM dendrimers have a particle size of approximately 4 nm [35]. Therefore, larger particles may reflect the formation of dendrimer agglomerates in PBS and populations near 4 and 5 nm are most likely related to dendrimers in free form. It is suggested in the literature that when the sample contains large aggregates, size determination of smaller particles by DLS can be compromised [36]. Because of this, we decided to evaluate PAMAM dispersion using another technique, i.e., Nanoparticle Tracking Analysis (NTA). NTA combines laser light scattering microscopy with a charge-coupled device (CCD) camera. This combination allows recording of the movement of individual particles when in a liquid because of Brownian motion. In this manner, the particles are identified and tracked, and their sizes are determined according to the Stokes-Einstein equation [37]. DLS and NTA are, therefore, both based on measurements of Brownian motion, and correlate particle motion with hydrodynamic diameter. The difference between them is that DLS measures scattering intensity fluctuations which are time-dependent, and calculates the particle size based on these fluctuations; whereas NTA measures the particle's Brownian motion by image analysis, attributing the particle movement to a specific particle size. NTA, therefore, provides the possibility of visualizing the dynamic of individual particles and the formation of aggregates. Table 1 shows zeta potentials and size distribution of PAMAM dispersions in PBS determined by NTA. The mean size of PAMAM G3.5 was slightly larger than PAMAM G4 but not significantly different (P N 0.05). PAMAM G3.5 exhibited a negative zeta potential as expected, and PAMAM G4 was slightly positive. NTA enabled us to visualize small particles that move rapidly, and the dynamic formation of large particles (Videos S1, S2, S3 and S4, supplemented material), which are assumed to be the PAMAM in free and aggregate forms, respectively. For all of the samples evaluated, the size distribution ranges from approximately 5 nm to 100 nm.

2.9. LC-MS analysis The amount of Dex in the samples prepared as described above was quantified using a Shimadzu HPLC System (Kyoto, Japan) equipped with two LC 20-AD solvent pumps, a CBM-20A system controller, a CTO-20A column oven and a LCMS-2020 mass spectrometer. Sample injections were made using a SIL-20AHT. The Dex was analyzed using a ShimPack VP-ODS (Shimadzu Corporation, Kyoto, Japan) C18 column (150 mm × 2.0 mm i.d., 5 μm) with a C18 column guard Shim-pack GVP-ODS (Shimadzu Corporation, Kyoto, Japan, 2.0 mm × 5 mm). The mobile phase was composed of 0.13% aqueous formic acid:ACN (70:30), and the flow rate was 0.2 mL/min. The method was developed based on defined conditions described in literature, with some modifications [31]. The injection volume was 25 μL, and the column oven temperature was programmed at 35 °C. The Dex in the samples was detected employing a Shimadzu LCMS 2020 mass spectrometer, equipped with a positive electrospray ion source. The detector parameters were established for single ion monitoring (SIM) of a Dex protonated molecular ion (M+ H+) at m/z 393.1. The analytical software used to control the system and acquire and process the data was LabSolutions (Ver. 5.42.30, Shimadzu Corporation). The method was validated for selectivity, sensitivity, linearity, accuracy and precision. The calibration curve was linear (r = 0.999) and the method was highly selective for Dex over the concentration range of 15–500 ng/mL.

3.1. Physicochemical properties of PAMAM G4 and PAMAM G3.5 in PBS

3.2. Effect of iontophoresis on PAMAM penetration and distribution in the cornea Fig. 1 shows the fluorescence distribution of transversal slices of the cornea after 3 h of passive and iontophoretic experiments with FITC, PAMAM G3.5-FITC and PAMAM G4-FITC. According to Fig. 1, passive application with FITC allowed a homogeneous and intense fluorescence distribution into the corneal epithelium. On the other hand, passive application with PAMAM G3.5-FITC and PAMAM G4-FITC leads to weak fluorescence intensities, primarily on

2.10. Data analysis Experiments were performed using four to five replicates. The data presented are shown as the mean ± standard deviation (for in vitro experiments) or the mean ± standard error of the mean (for ex vivo and in vivo experiments). The data were evaluated using an unpaired t-test (Prism, Graphpad Software, La Jolla, US). In all of the tests performed, P values b 0.05 were considered to be significantly different.

Table 1 Physicochemical properties of PAMAM G3.5 and PAMAM G4 in PBS, pH 7.4. Size was measured by NTA. Sample

Size (nm)

Zeta potential (mV)

PAMAM G3.5 PAMAM G4

54 (±16) 48 (±15)

−11.05 (±6.25) 5.69 (±0.44)

Data represent mean ± SD of 3 determinations.

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

119

Fig. 1. Photomicrographs of cornea transversal sections after 3 h of passive and iontophoretic permeation of FITC or PAMAM G3.5-FITC and PAMAM G4-FITC conjugates. A non-treated cornea was used as the control. All of the images were captured using the same parameters and were not processed further after capturing. Excitation/emission was performed at 480/ 515–530 nm. Scale bar = 250 μm.

the corneal surface. These results indicate that the fluorescence observed when PAMAM-FITC is delivered is due to FITC attached to PAMAM, corresponding, therefore, to PAMAM penetration. PAMAM does not considerably penetrate the corneal epithelium passively. The corneal epithelium is the outermost layer of the cornea and is the limiting factor for the absorption of hydrophilic drugs. Therefore, because dendrimers are large, hydrophilic molecules, low penetration into the epithelium was expected. Iontophoresis increased the intensity and depth of the fluorescence especially when FITC was attached to the dendrimers (Fig. 1), suggesting improved transport of the dendrimers across the epithelium toward the stroma. The epithelium has 5 cell layers [38], and the superficial cells present a large number of tight junctions [39]. Because the epithelium intercellular space is in the range of 2–3 nm [40,41], non-aggregated dendrimers could diffuse through the intercellular space because their size is similar to that of the intercellular pores, and also because of the semi-rigid container-type structure of generation 4–6 PAMAM dendrimers [35]. The transcellular pathway is another possible path by which dendrimers may enter into deep cornea layers by facilitated diffusion, active transport or endocytosis [39]. As PAMAM dendrimers have been shown to exhibit high uptake by several cell lines [19,42], the transcellular route may represent an important alternative for PAMAM transcorneal diffusion after iontophoresis translocates some of the molecules into the first corneal layer.

The PAMAM G3.5 and PAMAM G4 dispersions exhibited a negative and a positive zeta potential, respectively (Table 1). This result indicates that electromigration can play a role in PAMAM G4 penetration when in contact with the anode (anodal iontophoresis) and in PAMAM G3.5 penetration when in contact with the cathode (cathodal iontophoresis). However, the negative PAMAM G3.5 dispersion is not expected to take advantage of electromigration when in contact with the anode. In this case, electroosmotic flow from the anode to the cathode under physiological conditions [10] can pull PAMAM G3.5 into the cornea. In summary, the zeta potential of the dispersions suggests that PAMAM G4 can penetrate the cornea with the simultaneous contribution of electromigration and electroosmosis when in contact with the anode, and PAMAM G3.5 penetrates the cornea by electromigration when in contact with the cathode and by electroosmosis when in contact with the anode. To better evaluate the PAMAM distribution along the cornea after iontophoresis, optical sections parallel to the cornea surface were taken at different depths and periods of iontophoretic treatment (Fig. 2). Fig. 2 (left panel) shows that after 5 min of anodal iontophoresis, PAMAM G3.5 fluorescence was well observed to a depth of 20 μm; however, an intense fluorescence was observed to 35 μm when cathodal iontophoresis was performed. This result suggests that electromigration improves PAMAM G3.5 penetration more than electroosmosis and that the anodal iontophoresis of PAMAM G4 improved PAMAM penetration to approximately 25 μm.

120

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

Fig. 2. Optical sections parallel to the corneal surface at different depths after 5 min (left panel) and 3 h (right panel) of iontophoresis. A: PAMAM G3.5 anodal iontophoresis, B: PAMAM G3.5 cathodal iontophoresis and C: PAMAM G4 anodal iontophoresis. A non-treated cornea was used as the control. All of the images were captured using the same parameters and were not processed further after capturing. Excitation/emission was performed at 480/515–530 nm.

After 3 h of iontophoresis (Fig. 2, right panel), the fluorescence depth relative to both dendrimers after iontophoretic application was approximately 40 μm. According to the literature, the thickness of the corneal epithelium of Guinea pigs' eyes is 45.52 ± 5.26 μm [43]. Therefore, it appears that iontophoresis improved dendrimer transport across the entire epithelium thickness to the beginning of the next corneal layer, the stroma. The distribution of dendrimers in this layer appears to be different for PAMAM G4 and PAMAM G3.5. The cationic PAMAM G4 (Fig. 2C, right panel) exhibits a more homogeneous distribution than the anionic PAMAM G3.5. Cationic PAMAMs exhibited a higher transepithelial transport than anionic PAMAM in Caco-2 [26], which can explain the more homogenous distribution of PAMAM G4 in deep cornea cells after iontophoresis. A control experiment with free FITC was performed for 3 h under the same conditions. The results suggest that FITC penetrates less into the corneal epithelium under iontophoretic application than PAMAM-FITC conjugates (Supplemental material, Figure S1). Moreover, the FITC fluorescence distribution profile was different from PAMAM-FITC, suggesting again that the fluorescence seen when PAMAM-FITC was administered was due to FITC attached to the dendrimer and, therefore, to PAMAM distribution. 3.3. Dex-PAMAM complexes characterization After characterization of PAMAM dendrimers and the suggestion that they can penetrate the cornea under the influence of iontophoresis, the dendrimers were complexed with Dex to evaluate the influence of PAMAM on the penetration of Dex into the cornea. Complex formation between Dex and PAMAM was demonstrated using different NMR techniques (Figures S2 to S7 in supplemental material) and solubility studies. In the 1H-NMR and 13C-NMR spectra of the complex, signals for dendrimers and Dex could be observed, which suggests complex

formation. Moreover, DOSY analyses demonstrated a decrease of the diffusion coefficient for both PAMAMs after complexation with Dex, which is explained in detail in the supplementary material. The latter results were also consistent with the previous 1H-NMR and 13C-NMR conclusions. PAMAM complexation with Dex did not significantly change the mean particle size for both dendrimers or their zeta potential (Table 1), which were 53 ± 18 nm and − 17.4 ± 1.27 mV for DexPAMAM G3.5, and 45 ± 14 nm and 5.13 ± 0.51 mV mV for DexPAMAM-G4 in PBS, pH 7.4 (particle size distribution in supplemental material, Figure S8). These characteristics suggest that the main interactions between Dex and PAMAMs may occur within the PAMAM hydrophobic cavities because possible interactions with PAMAM terminal groups could have changed the zeta potential. The effect of each PAMAM on the Dex water solubility was investigated, yielding values of 91.67, 353.15 and 939.80 μg/mL for Dex, DexPAMAM G4 and Dex-PAMAM G3.5, respectively. For a lipophilic drug such as Dex, an increase in aqueous solubility should improve both cell uptake [20] and drug bioavailability [22,24]. According to the results, complexation with PAMAM G4 increased the solubility of Dex in water 4-fold, while complexation with PAMAM G3.5 resulted in a 10-fold increase. These results are consistent with those of other studies of solubility with a lipophilic drug (nifedipine) and PAMAM dendrimers. Researchers have demonstrated that PAMAM dendrimers of “half generation” increase nifedipine water solubility more than the full generation PAMAMs [21]. It appears that at physiological pH, half-generation dendrimers have a few internal tertiary amines protonated compared with PAMAM G4. Consequently, many non-protonated internal amines are available for hydrogen bonding with the lipophilic drug [44]. In addition to the possible hydrogen bond interactions, the non-protonated tertiary amines could provide greater hydrophobicity to the PAMAM G3.5 internal cavities, promoting stronger interactions with these

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

cavities and the lipophilic Dex. As a result, the interactions between Dex and these internal cavities are stronger for PAMAM G3.5 than for PAMAM G4, accommodating more Dex molecules and further increasing the aqueous solubility of the drug. The osmolalities of Dex-PAMAM G3.5 and Dex-PAMAM G4 complexes in phosphate buffer, pH 7.4 were 289 and 291 mOsm/Kg, respectively. These values are in agreement with the limits described in the literature for ophthalmic formulations (between 260 and 340 mOsm/ Kg), thereby allowing it to be safe for clinical applications [45,46].

121

Table 2 Dex release rate (K) from various Dex saturated formulations in PBS, pH 7 and the Dex diffusion coefficient (D). PBS

K (μg/cm2/h1/2)

D × 10−7 (cm/s)

Dex Dex-PAMAM G4 Dex-PAMAM G3.5

14.1 (±1.1) 39.7 (±5.3)* 14.4 (±3.1)

33.0 (±7.5) 19.4 (±4.9)*# 0.4 (±0.2)*

Data represent the mean ± SD of 4–5 determinations. * indicates a significant difference compared to Dex (Unpaired t test, P b 0.05). # indicates significant difference compared to Dex-PAMAM G3.5 (Unpaired t test, P b 0.05).

3.4. In vitro release studies Fig. 3 shows the amount and the percentage of Dex released as a function of time from each of the saturated dispersion. The amount of drug released from Dex-PAMAM G3.5 was similar to the amount released by free Dex dispersion. Therefore, as free Dex has been successfully used clinically in the form of eye drops, Dex-PAMAM G3.5 is not expected to decrease drug efficacy. The amount of Dex released from the complex was observed to be linear when plotted against the square root of time (r N 0.98). The value of K, given in Table 2, was higher for the Dex-PAMAM G4 complex, and almost 3-fold greater than that for the Dex solution and the DexPAMAM G3.5 complex. However, considering the diffusion coefficient (D), i.e., the amount of drug present in the donor dispersion, complexation with PAMAMs decreased the D for Dex. The decrease was especially impressive for the Dex-PAMAM G3.5 formulation (more than 90-fold less) (Table 2 and Fig. 3B). These results indicate that PAMAM G3.5 may interact strongly with the lipophilic Dex as suggested by the solubility studies, thus encapsulating more Dex but also delaying the drug release. 3.5. Dex-PAMAM ex vivo cornea permeation studies Ex vivo cornea permeation studies were conducted to investigate whether the dendrimers were able to release adequate drug concentrations into and through the cornea. Therefore, Dex permeation was evaluated for a period of 3 h. Although this application time is not feasible for clinical applications it has been chosen to ensure the drug quantification in the ex vivo studies and to enable comparisons with the Dex administered in the absence of dendrimer. Dex is usually administered as a suspension at a concentration of 1 mg/mL (e.g., Maxidex®, Maxitrol®). Therefore, this was the concentration used in the ex vivo experiments.

The Dex concentrations for Dex-PAMAM G3.5 and Dex-PAMAM G4 complexes were maintained in saturated conditions at 939.80 μg/mL and 353.15 μg/mL, respectively. In other words, each complex contains the maximum Dex concentration so as to maintain the equilibrium between Dex and PAMAM and guarantee high complexation degree. Therefore, penetration of Dex from Dex-PAMAM saturated solutions was compared with penetration of Dex from a suspension at 1 mg/mL. In all cases, the Dex permeation rate followed zero-order kinetics, i.e., it was independent of the drug concentration in the donor. Table 3 lists the Dex accumulation in the receptor solution after 3 h of passive and iontophoretic experiments. According to Table 3, Dex-PAMAM G4 passive application enhanced the amount of Dex that permeated by almost 2.2-fold (P 0.05), while Dex-PAMAM G3.5 did not alter the Dex permeation significantly. As the amount of Dex in the donor for the Dex-PAMAM G4 dispersion was smaller than that of other dispersions, it is suggested that PAMAM G4 modifies the corneal surface to some extent, facilitating drug permeation. Indeed, investigations of the interactions between cationic PAMAMs (G4 and G5) and lipid bilayer membranes have revealed that these dendrimers interact with the membrane, modifying its organization [47,48]. It appears, therefore, that PAMAM G4 modifies the corneal surface to some extent, facilitating Dex permeation. Iontophoresis increased Dex permeation in all cases compared with the passive experiments, but to a different extent for each formulation (Table 3). Anodal iontophoresis enhanced the amount of drug permeated 2.9, 5.6 and 3.0-fold (P b 0.05) for Dex suspension, Dex-PAMAM G4 and Dex-PAMAM G3.5, respectively. The amount of Dex delivered through the cornea after 3 h of cathodal and anodal iontophoresis of Dex-PAMAM G3.5 was not significantly different (P b 0.05). Therefore, it was not possible to observe differences in Dex permeation between electromigratory and electroosmotic transport of the dendrimers. This

Fig. 3. In vitro Dex release from saturated Dex, Dex-PAMAM G4 and Dex-PAMAM G3.5 dispersions as a function of time. A) Amount released per area; B) Percentage released as a function of donor concentration. Data represent mean ± SD (n = 4–5).

122

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

Table 3 Dex accumulation in the receptor solution after 3 h of passive and iontophoretic experiments. The Dex initial concentrations in the donor solutions were 1000 μg/mL, 939.80 μg/mL and 353.15 μg/mL of Dex, Dex-PAMAM G3.5 and Dex-PAMAM G4, respectively. The values represent the mean ± SEM of 4–5 determinations. Different symbols represent statistical differences (Unpaired t test, P b 0.05). Drug in PBS

Dex Dex-PAMAM G4 Dex-PAMAM G3.5

Dex accumulation after 3 h (μg/cm2) Passive

Anodal iontophoresis

Cathodal iontophoresis

0.73 (±0.14)⁎ 1.62 (±0.22)⁎⁎ 0.70 (±0.07)⁎

2.12 (±0.14)⁎⁎ 9.13 (±1.07)& 2.13 (±0.53)⁎⁎

– – 1.54 (±0.26)⁎⁎

result corroborates the confocal microscopy results of the cornea in the presence of dendrimers, which shows no difference between the PAMAM G3.5 cathodal and anodal distribution over the corneal cells after a long period of iontophoresis (Fig. 2, right panel). The remarkable increase in Dex permeation induced by the iontophoresis of the Dex-PAMAM G4 complex, despite the low concentration of Dex in the donor solution (3-fold smaller than that in the noncomplexed Dex suspension), indicates that PAMAM G4 must have facilitated corneal permeation of Dex. Therefore, the high iontophoretic permeation of Dex from this dendrimer suggests the penetration of the dendrimer into the cornea, corroborating the experiments performed with labeled PAMAM (Figs. 1 and 2). Although PAMAM G3.5 also penetrates the cornea when iontophoresis was applied (Figs. 1 and 2), the low Dex diffusion coefficient from this complex (Table 2) likely affects the amount of Dex that permeated the cornea. In addition to Dex permeation, at the end of the experiments, Dex was recovered from the corneas and quantified by LC-MS. Fig. 4 shows the amount of Dex recovered from the cornea after passive and iontophoretic experiments with Dex, Dex-PAMAM G3.5 and Dex-PAMAM G4 complexes dispersions in PBS. The Dex cornea recovered was twice as large for all formulations studied after iontophoresis compared with the passive experiments. The Dex-PAMAM G4 formulation transported more Dex into the cornea; this result is consistent with the permeation studies discussed previously. Once again, this suggests that PAMAM G4 may have changed the cornea, improving penetration of the complex. Additionally, because the interactions between Dex and this dendrimer are weaker than those between Dex and PAMAM G3.5, Dex is released faster (Table 2), resulting in a greater amount of Dex in the receiver compartment than from Dex-PAMAM G3.5 (Table 3).

Fig. 4. Dex uptake into the cornea after 3 h of ex vivo passive and iontophoretic permeation experiments from PBS dispersions at pH 7.4 containing Dex free or complexed with PAMAM G3.5 or G4 dendrimers. The current density applied was 1.0 mA/cm2. The data presented are the mean ± SD of 4–5 replicates (* represents statistically significant differences between iontophoresis and passive application for each formulation. Unpaired t test, P b 0.05).

3.6. Dex-PAMAM in vivo cornea permeation The main purpose of this first in vivo study was to determine if the iontophoresis-PAMAM association could provide higher drug concentrations in the aqueous humor compared to the administration of free drug or to the complexes applied passively. Iontophoresis was applied for only 4 min, and formulations were placed in contact with the electrode of opposite polarity in relation to their zeta potential to obtain the contribution of electromigration. Additionally, total electric current density applied was 20.4 mA/cm2 (5.1 mA/cm2 for 4 min), instead of the 180 mA/cm2 (1 mA/cm2 for 3 h) used in ex vivo experiments. In general, an increase in the electric current density and the duration of the application enhances drug penetration [30,49]. Therefore, the ex vivo experiments were performed for an extended period of time to ensure quantification of the drug, but extended periods of application are not comfortable and are not tolerated in vivo. Despite the safety of the association between Dex-PAMAM complexes and iontophoresis for transcorneal drug delivery has not been evaluated in this study, we used the same electric current density of other authors in transcorneal iontophoresis in vivo studies [30]. Fig. 5 shows that Dex concentrations recovered from the aqueous humor after 2 h of formulation application. According to Fig. 4, there was no difference in the Dex concentration in the aqueous humor when formulations were applied passively (concentrations between 34 and 41 ng/mL). Iontophoresis enhanced the amount of drug that permeated for all formulations. For Dex and DexPAMAM G4, there was an increase of 2 and 2.5-fold, respectively. Iontophoresis of Dex-PAMAM G3.5 increased Dex permeation more that 6fold in vivo, despite the smaller increment, compared to Dex-PAMAM G4, observed in the ex vivo experiments (Table 3). It is important to consider that iontophoresis was applied for 3 h in the ex vivo experiments, while only 4 min application was performed in the in vivo ones. Therefore, the expected permeation enhancer effect of PAMAM G4 suggested in the ex vivo experiments was not observed in vivo. It seems that the protective mechanisms of the eyes such as lacrimation and tear turnover caused dilution and drainage of the complexes. A short-time iontophoretic application however was able to push some PAMAM dendrimers into the cornea as observed in Fig. 2 (left panel). Therefore, the higher amount of Dex found in the aqueous humor after Dex-PAMAM G3.5 iontophoresis may be related to the ability of PAMAM G3.5 to sustain Dex release (Table 2 and Fig. 3). In the ex vivo studies, the drug that permeates the cornea and reaches the receiver solution was not under the influence of factors found in the in vivo experiments, such as enzymatic degradation and aqueous humor renewal. Once the drug has reached the anterior chamber, it is removed from this area through the trabecular meshwork and Sclemm's canal [50]. Therefore, as PAMAM G3.5 contains a greater amount of Dex, once Dex-PAMAM G3.5 complex has penetrated into the most superficial corneal it may provide sustained drug release, resulting in a higher

Fig. 5. Dex recovered from aqueous humor after 2 h of passive and 4 min iontophoretic in vivo application of Dex, Dex-PAMAM G4 and Dex-PAMAM G3.5. Values represent the mean ± SEM of 4 determinations (Unpaired t test, *P b 0.05).

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124

drug concentration in the anterior chamber even with dynamic drug removal. On the other hand, the Dex-PAMAM G4 complex carries fewer drug molecules, and drug-dendrimer interactions are weaker compared to Dex-PAMAM G3.5, causing a more rapid release of drug from the complex and faster elimination from the anterior chamber. The more significant performance of iontophoresis of PAMAM G3.5 complex is also relevant when the toxicity of the dendrimers is taken into consideration. PAMAM toxicity is dependent on the concentration, the generation and the number/nature of surface groups [26]. In cultures of Caco-2 cells, PAMAM G3.5 kills 10% of the cells, but only when applied at a concentration of 1000 μmol/L. On the other hand, at a concentration of 100 μmol/L, the cationic PAMAM G4 exhibited a toxicity of approximately 50% [26]. Therefore, our results demonstrate that the less cytotoxic PAMAM G3.5 dendrimer provided greater drug bioavailability in the in vivo experiments, encouraging future investigations with this system. 4. Conclusions Our findings demonstrate that 4th generation PAMAM dendrimers dispersed in isosmotic PBS yielded 5 nm particles that move rapidly leading to a dynamic formation of aggregates of approximately 50 nm. Iontophoresis was able to transport PAMAM G3.5 and G4 deep into the cornea. The distribution of the dendrimers throughout the cornea was dependent on the iontophoretic application time. Complexing these dendrimers with Dex significantly increased the water solubility of Dex, although the size and zeta potentials of the PAMAMs remained unchanged. The in vitro release studies demonstrated that PAMAM G3.5 decreased the Dex diffusion coefficient 53-fold, compared with PAMAM G4. Ex vivo experiments showed that Dex-PAMAM G4 complex enhanced Dex cornea permeation by approximately 2 and 4-fold by passive and iontophoretic delivery, respectively, compared with Dex and Dex-PAMAM G3.5. However, in vivo experiments have shown that iontophoresis of PAMAM G3.5 leads to almost 3-fold greater amounts of Dex in the aqueous humor than PAMAM G4. In vivo iontophoresis increased the Dex concentration in the aqueous humor by 2, 2.5 and 6.6-fold for Dex, Dex-PAMAM G4 and Dex-PAMAM G3.5, respectively, compared to passive applications. In summary, this work has shown for the first time the potential of iontophoresis in non-invasive corneal administration of dendrimers. Under the conditions in which the experiments were performed, it can be concluded that iontophoresis enhances the penetration of PAMAM dendrimers of 4th generation through the cornea. Moreover, it appears that interactions between the drug and the dendrimer must be sufficiently strong to sustain drug release under physiological conditions for dendrimer penetration to have an impact on corneal drug bioavailability. Therefore, the combination of iontophoresis and PAMAM dendrimer represents an interesting tool that can be applied toward topical and sustained drug delivery for the treatment of eye diseases. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2014.12.037. Acknowledgments The authors would like to thank the São Paulo Research Foundation (FAPESP grant #11/11305-1 and #10/19210-7), the Brazilian National Council for Scientific and Technological Development (CNPq grant #565361/2008-2 and #480962/2013-8) and the Core Research in Pathophysiology and Ocular Therapeutics (NAP-FTO), University of São Paulo, grant # 12.1.25431.01.7, Brazil for financial support. References [1] R.E. Grossman, D.F. Chu, D.A. Lee, Regional ocular gentamicin levels after transcorneal and transscleral iontophoresis, Invest. Ophthalmol. Vis. Sci. 31 (1990) 909–916.

123

[2] D. Monti, L. Saccomani, P. Chetoni, S. Burgalassi, M.F. Saettone, Effect of iontophoresis on transcorneal permeation ‘in vitro’ of two beta-blocking agents, and on corneal hydration, Int. J. Pharm. 16 (2003) 423–429. [3] T. Gratieri, G.M. Gelfuso, J.A. Tomazini, R.F. Lopez, Excised porcine cornea integrity evaluation in an in vitro model of iontophoretic ocular research, Ophthalmic Res. 43 (2010) 208–216. [4] S.A. Molokhia, E.K. Jeong, W.I. Higuchi, S.K. Li, Transscleral iontophoretic and intravitreal delivery of a macromolecule: study of ocular distribution in vivo and postmortem with MRI, Exp. Eye Res. 88 (2009) 418–425. [5] S. Pescina, G. Ferrari, P. Govoni, C. Macaluso, C. Padula, P. Santi, S. Nicoli, In-vitro permeation of bevacizumab through human sclera: effect of iontophoresis application, J. Pharm. Pharmacol. 62 (2010) 1189–1194. [6] L. Von Sallmann, K. Meyer, Penetration of penicillin into the eye, Arch. Ophthalmol. 31 (1944) 1–7. [7] T. Asahara, K. Shinomiya, T. Naito, H. Shiota, Induction of gene into the rabbit eye by iontophoresis: preliminary report, Jpn. J. Ophthalmol. 45 (2001) 31–39. [8] M. Berdugo, F. Valamanesh, C. Andrieu, C. Klein, D. Beneza, Y. Courtois, F. BeharCohen, Delivery of antisense oligonucleotide to the cornea by iontophoresis, Antisense Nucleic Acid Drug Dev. 13 (2003) 107–114. [9] J. Hao, S.K. Li, C.Y. Liu, W.W. Kao, Electrically assisted delivery of macromolecules into the corneal epithelium, Exp. Eye Res. 86 (2009) 934–941. [10] D. Marro, Y.N. Kalia, M.B. Delgado-Charoo, R.H. Guy, Contributions of electromigration and electroosmosis to iontophoretic drug delivery, Pharm. Res. 18 (2001) 1701–1708. [11] M.A. Patane, W. Schubert, T. Sanford, R. Gee, M. Burgos, W.P. Isom, B. Ruiz-Perez, Evaluation of ocular and general safety following repeated dosing of dexamethasone phosphate delivered by transscleral iontophoresis in rabbits, J. Ocul. Pharmacol. Ther. 29 (2013) 760–769. [12] S. Güngör, M.B. Delgado-Charro, B. Ruiz-Perez, W. Schubert, P. Isom, P. Moslemy, M.A. Patane, R.H. Guy, Trans-scleral iontophoretic delivery of low molecular weight therapeutics, J. Control. Release 47 (2010) 225–231. [13] M.A. Patane, A. Cohen, S. From, G. Torkildsen, D. Welch, G.W. Ousler, Ocular iontophoresis of EGP-437 (dexamethasone phosphate) in dry eye patients: results of a randomized clinical trial, Clin. Ophthalmol. 5 (2011) 633–643. [14] P. Chopra, J. Hao, S.K. Li, Sustained release micellar carrier systems for iontophoretic transport of dexamethasone across human sclera, J. Control. Release 160 (2012) 96–104. [15] A.E. Cohen, C. Assang, M.A. Patane, S. From, M. Korenfeld, Evaluation of dexamethasone phosphate delivered by ocular iontophoresis for treating noninfectious anterior uveitis, Ophthalmology 119 (2012) 66–73. [16] E. Eljarrat-Binstock, F. Orucov, Y. Aldouby, J. Frucht-Pery, A.J. Domb, Charged nanoparticles delivery to the eye using hydrogel iontophoresis, J. Control. Release 126 (2008) 156–561. [17] T.F. Vandamme, L. Brobeck, Poly(amidoamine) dendrimers as ophthalmic vehicles for ocular delivery of pilocarpine nitrate and tropicamide, J. Control. Release 102 (2005) 23–38. [18] M. El-Sayed, M. Ginski, C.A. Rhodes, H. Ghandehari, Transepithelial transport of poly(amidoamine) dendrimers on Caco-2 cell monolayers, J. Control. Release 81 (2002) 151–157. [19] O.P. Perumal, R. Inapagolla, S. Kannan, R.M. Kannan, The effect of surface functionality on cellular trafficking of dendrimers, Biomaterials 29 (2008) 3469–3476. [20] P. Kolhe, E. Misra, R.M. Kannan, S. Kannan, M. Lieh-Lai, Drug complexation, in vitro release and cellular entry of dendrimers and hyperbranched polymers, Int. J. Pharm. 259 (2003) 143–160. [21] B. Devarakonda, R.A. Hill, M.M. Villiers, The effect of PAMAM dendrimer generation size and surface functional group on the aqueous solubility of nifedipine, Int. J. Pharm. 284 (2004) 133–140. [22] M. Na, C. Yiyun, X. Tongwen, D. Yang, W. Xiaomin, L. Zhenwei, C. Zhichao, H. Guanyi, S. Yunyu, W. Longping, Dendrimers as potential drug carriers. Part II. Prolonged delivery of ketoprofen by in vitro and in vivo studies, Eur. J. Med. Chem. 41 (2006) 670–674. [23] Y.E. Kurtoglu, R.S. Navath, B. Wang, S. Kannan, R. Romero, R.M. Kannan, Poly(amidoamine) dendrimer–drug conjugates with disulfide linkages for intracellular drug delivery, Biomaterials 30 (2009) 2112–2121. [24] W. Yao, K. Sun, H. Mu, N. Liang, Y. Liu, C. Yao, R. Liang, A. Wang, Preparation and characterization of puerarin–dendrimer complexes as an ocular drug delivery system, Drug Dev. Ind. Pharm. 36 (2010) 1027–1035. [25] W.Y. Wang, C. Yao, Y.F. Shao, H.J. Mu, K.X. Sun, Determination of puerarin in rabbit aqueous humor by liquid chromatography tandem massspectrometry using microdialysis sampling after topical administration of puerarin PAMAM dendrimercomplex, J. Pharm. Biomed. Anal. 56 (2011) 825–829. [26] R. Jevprasesphant, J. Penny, D. Attwood, N.B. McKeown, A. D'Emanuele, Engineering of dendrimer surfaces to enhance transepithelial transport and reduce cytotoxicity, Pharm. Res. 20 (2003) 1543–1550. [27] R. Wiwattanapatapee, B. Carreno-Gomez, N. Malik, R. Duncan, Anionic PAMAM dendrimers rapidly cross adult rat intestine in vitro: a potential oral delivery system, Pharm. Res. 17 (2000) 991–998. [28] T. Gratieri, G.M. Gelfuso, O. de Freitas, E.M. Rocha, R.F. Lopez, Enhancing and sustaining the topical ocular delivery of fluconazole using chitosan solution and poloxamer/chitosan in situ forming gel, Eur. J. Pharm. Biopharm. 79 (2011) 320–327. [29] W.I. Higuchi, Analysis of data on medicament release from ointments, J. Pharm. Sci. 51 (1962) 802–804. [30] E. Eljarrat-Binstock, F. Raiskup, J. Frucht-Pery, A.J. Domb, Transcorneal and transscleral iontophoresis of dexamethasone phosphate using drug loaded hydrogel, J. Control. Release 106 (2005) 386–390. [31] P. Patel, S. Tanna, H. Mulla, V. Kairamkonda, H. Pandya, G. Lawson, Dexamethasone quantification in dried blood spot samples using LC-MS: the potential for application

124

[32]

[33] [34]

[35] [36] [37] [38] [39] [40]

[41] [42]

J.G. Souza et al. / Journal of Controlled Release 200 (2015) 115–124 to neonatal pharmacokinetic studies, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 878 (2010) 3277–3282. V. Filipe, A. Hawe, W. Jiskoot, Critical evaluation of Nanoparticle Tracking Analysis (NTA) by NanoSight for the measurement of nanoparticles and protein aggregates, Pharm. Res. 27 (2010) 796–810. S. Khouri, K.C. Tam, Complexation between α-cyclodextrin and PEGylated-PAMAM dendrimers at low and high pH values, Langmuir 26 (2010) 17969–17974. S. Sadekar, H. Ghandehari, Transepithelial transport and toxicity of PAMAM dendrimers: implications for oral drug delivery, Adv. Drug Deliv. Rev. 64 (2012) 571–588. R. Esfand, D.A. Tomalia, Poly(amidoamine) (PAMAM) dendrimers: from biomimicry to drug delivery and biomedical applications, Drug Discov. Today 6 (2001) 427–436. B. Berne, R. Pecora, Dynamic light scattering with applications to chemistry, biology, and physics, Mineola, Dover, 2000. http://nanosight.com.uk (Accessed in February 18, 2014). J.S. Pepose, J.L. Ubels, The cornea, in: W.M. Hart Jr. (Ed.), Adler's Physiology of the Eye, Mosby-Year Book, Inc., St. Louis, 1992, pp. 29–70. G. Sunkara, U.B. Kompella, Membrane transport processes in the eye, in: A.K. Mitra (Ed.), Ophthalmic Drug Delivery Systems, Marker Decker, New York, 2003, pp. 13–58. K.M. Hämäläinen, K. Kontturi, S. Auriola, L. Murtomäki, A. Urtti, Estimation of pore size and pore density of biomembranes from permeability measurements of polyethylene glycols using an effusion-like approach, J. Control. Release 49 (1997) 97–104. V.H.L. Lee, L.W. Carson, K.A. Takemoto, Macromolecular drug absorption in the albino rabbit eye, Int. J. Pharm. 29 (1986) 43–51. H. Dai, R.S. Navath, B. Balakrishnan, B.R. Guru, M.K. Mishra, R. Romero, R.M. Kannan, S. Kannan, Intrinsic targeting of inflammatory cells in the brain by polyamidoamine

[43]

[44]

[45] [46]

[47]

[48] [49]

[50]

dendrimers upon subarachnoid administration, Nanomedicine (Lond.) 9 (2010) 1317–1329. T.A. Cafaro, S.G. Ortiz, C. Maldonado, F.A. Espósito, J.O. Croxatto, A. Berra, O.L. Ale, J.I. Torrealday, F.A. Urrets-Zavalía, J.A. Urrets-Zavalía, H.M. Serra, The cornea of Guinea pig: structural and functional studies, Vet. Ophthalmol. 12 (2009) 234–241. M.F. Ottaviani, F. Montalti, M. Romanelli, N.J. Turro, D.A. Tomalia, Characterization of starburst dendrimers by EPR. 4. Mn(II) as a probe of interphase properties, J. Phys. Chem. 100 (1996) 11033–11042. A. Ludwig, M. Van Ooteghem, The influence of the osmolality on the precorneal retention of ophthalmic solutions, J. Pharm. Belg. 42 (1987) 259–266. L.M. Bowman, E. Si, J. Pang, R. Archibald, M. Friedlaender, Development of a topic polymeric mucoadhesive ocular delivery system for azithromycin, J. Ocul. Pharmacol. Ther. 25 (2009) 133–139. K. Gardikis, S. Hatziantoniou, K. Viras, M. Wagner, C. Demetzos, A DSC and Raman spectroscopy study on the effect of PAMAM dendrimer on DPPC model lipid membranes, Int. J. Pharm. 318 (2006) 118–123. D. Shcharbin, A. Drapeza, V. Loban, A. Lisichenok, M. Bryszewska, The breakdown of bilayer lipid membranes by dendrimers, Cell. Mol. Biol. Lett. 11 (2006) 242–248. F.F. Behar-Cohen, A. El Aouni, S. Gautier, G. David, J. Davis, P. Chapon, J.M. Parel, Transscleral Coulomb-controlled iontophoresis of methylprednisolone into the rabbit eye: influence of duration of treatment, current intensity and drug concentration on ocular tissue and fluid levels, Exp. Eye Res. 74 (2002) 51–59. A. URTTI, Challenges and obstacles of ocular pharmacokinetics and drug delivery, Adv. Drug Deliv. Rev. 58 (2006) 1131–1135.