Ocular penetration of fluorometholone-loaded PEG-PLGA nanoparticles functionalized with cell-penetrating peptides

Roberto Gonzalez-Pizarro1,5, Graziella Parrotta3, Rodrigo Vera2,6,7, Elena Sánchez-López1,4,5, Ruth Galindo1, Frank Kjeldsen3, Josefa Badia2,6,7, Laura Baldoma2,6,7, Marta Espina1,5 & Marı́a L Garcı́a*,1,4,5 1Department of Pharmacy, Pharmaceutical Technology & Physical Chemistry, Faculty of Pharmacy & Food Sciences, University of Barcelona, 08028 Catalonia, Spain 2Department of Biochemistry & Physiology, Faculty of Pharmacy & Food Sciences, University of Barcelona, 08028 Catalonia, Spain 3Department of Biochemistry & Molecular Biology, Faculty of Science, University of Southern Denmark, 5230 Southern Denmark, Denmark 4Biomedical Research Networking Center in Neurodegenerative Diseases (CIBERNED), 28031 Madrid, Spain 5Institute of Nanoscience & Nanotechnology (IN2UB), University of Barcelona, 08028 Catalonia, Spain 6Institute of Biomedicine, University of Barcelona (IBUB), 08028 Catalonia, Spain 7Institut de Recerca Sant Joan de Déu (IRSJD), 08950 Catalonia, Spain *Author for correspondence: marisagarcia@ub.edu

inactivation by cytochrome P450, and its elimination by efflux transports and multidrug resistance proteins [15,16]. Therefore, the new nanostructured systems should protect the drug and allow internalization in ocular cells avoiding its elimination by both molecular (metabolizing enzymes) and anatomical (tear turnover and dilution) mechanisms. With regard to the previous approach, cell-penetrating peptides (CPPs) constitute a new strategy that would allow the internalization of nanostructured systems such as NPs. CPPs are sequences of less than 30 amino acids, rich in arginine (Arg) residues and positively charged. The use of CPPs on the surface of the NPs improve their penetration through cell membranes using various mechanisms such as direct translocation or receptor-mediated endocytosis. This allows a greater pharmacological effect of the encapsulated component than NPs without CPPs on their surface [17]. CPPs such as Penetratin (pAntp [43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58] ) and the transcriptional activator peptide sequence (TAT 49-57 ) of HIV-1 have been shown to be effective for the cellular internalization of polypeptides and the transfection of nucleic acids. Besides, many of the CPPs share characteristics with antimicrobial peptides, such as cationic sequences with multiple Arg residues [18][19][20]. Within the antimicrobial peptide group, the peptide sequence G2 has antiviral activity against the herpes simplex virus and has been shown to have the ability to internalize drugs inside cells [21][22][23][24][25].

Synthesis of the polymers Rho-PLGA & m-PEG-PLGA
A total of 33.3 μmol of PLGA RG 503H (1.0 g) were dissolved in 3 ml of chloroform, then allowed to react with 234.6 μmol of NHS (27 mg) and EDC (45 mg) overnight under continuous stirring in a sealed glass vial. Subsequently, PLGA RG 503 H (NHS-PLGA) was activated, precipitated with 10 ml of cold diethyl ether and centrifuged at 4000 r.p.m for 10 min at 20 • C. The supernatant was removed, the recovered polymer was dissolved in 3 ml of chloroform and precipitated with cold diethyl ether. This washing/precipitation cycle was carried out three times. The NHS-PLGA obtained was dried using nitrogen gas (N 2 ) and then lyophilized and stored at −20 • C.
A total of 33.2 μmol of NHS-PLGA (1.0 g), 35 μmol of m-PEG (167 mg) or Rho (12.78 mg) were dissolved in 3 ml of chloroform with 234.3 μmol of DIEA and left to react overnight under continuous stirring. The obtained polymer was subjected to three washing/precipitation cycles using an 80/20 mixture of diethyl ether/cold methanol. At last, the polymer was lyophilized and stored at −20 • C. The yield obtained from polymers (m-PEG-PLGA and Rho-PLGA) was calculated by the following equation: weight of (m-PEG-PLGA) weight of (NHS-PLGA + m-PEG) × 100 (Eq. 1) To follow the synthesis of the polymer and changes in its molecular structure, it was analyzed by H-NMR. PLGA RG 503H, NHS-PLGA, m-PEG, m-PEG-PLGA and Rho-PLGA were dissolved in deuterated dimethyl sulfoxide (DMSO-d 6 ). The spectrum was recorded at 298 K on a Varian VNMRS 400 MHZ spectrometer (Agilent Technologies, CA, USA). To calculate the PEGylation efficiency and percentage of mass relative (MR) of the m-PEG bound to the NHS-PLGA polymer, the following equations were used [17,26]

Structural studies of polymers
The molecular changes of the synthesized polymer and its crystalline or amorphous state were evaluated by FTIR and XRD. FTIR spectra of polymers separately were obtained using a Thermo Fisher Scientific Nicolet iZ10 with an ATR diamond and DTGS detector. The scanning range was 525-4000 cm −1 [27]. The amorphous or crystalline state of NHS-PLGA, m-PEG and m-PEG-PLGA were determined by XRD measurements using Siemens D500 system (Karlsruher, Germany). X-ray powder diffractograms were recorded using Cu K radiation (45 kV, 40 mA, λ = 1.544Å) in the range (2θ) from 2 • to 60 • with a step size of 0.026 • and measuring time of 195.8 s per step [27].
Preparation & characterization of FMT-m-PEG-PLGA-NPs FMT-m-PEG-PLGA-NPs and fluorescent NPs ( Rho NPs) were prepared by the solvent displacement method [27]. Briefly, the m-PEG-PLGA (10.0 mg·ml −1 ) and FMT (0.5 mg·ml −1 ) were dissolved in 5 ml of acetone. The organic phase was added slowly dropwise, under stirring, into 10 ml of an aqueous solution of PVA (15 mg·ml −1 ) adjusted to pH 4.0. Subsequently, the organic solvent was evaporated under reduced pressure. At last, the PVA was eliminated by ultracentrifugation (30,000 r.p.m for 30 min) and resuspended in 10 ml of phosphate buffered saline (PBS). The same amounts of polymer (20 mg of Rho-PLGA and 80 mg of m-PEG-PLGA) and FMT were used for the synthesis of Rho NPs.
Morphometry (average particle size [Zav] and polydispersity index [PI] measured from scattering intensity) and ζ-potential (ZP) of NPs were determined in a Zetasizer NanoZS (Malvern Instruments, Malvern, UK) by dynamic light scattering and electrophoretic mobility, respectively. Samples were diluted in MilliQ water (1:10) and experiments were performed with disposable capillary cells DTS1070 (Malvern Instruments) at 25 • C. The reported values correspond to the mean ± SD.
The entrapment efficiency (EE) of FMT in the NPs was quantified indirectly by measuring the nonentrapped drug in the dispersion medium. The free FMT was separated by a filtration/centrifugation technique (1:10 dilution) at 25 • C and 5000 r.p.m for 10 min. The EE was calculated according to the following equation: where cFMT 0 and cFMT 1 are the total amount of FMT and free FMT in the filtrated, respectively. Samples were evaluated by HPLC, according to the method described previously [27]. Data were processed using Empower 3 R software.

Cytotoxicity study
The effect on cell viability of NPs conjugated with each CPP was evaluated in vitro using the MTT assay. For this, 1 × 10 5 HCE-2 cells in 0.1 ml were seeded in 96-well plates and incubated for 24 h at 37 • C. Then, cells were exposed to different drug concentrations contained in TAT-NPs, pAntp-NPs, G2-NPs and Isoptoflucon (ophthalmic commercial drug). After 24 or 48 h of incubation, cells were washed with PBS and incubated with 0.25 % MTT in fresh medium for 2 h at 37 • C in dark. Subsequently, medium was removed, and cells were lysed by the addition of 99% DMSO. At last, the absorbance was measured at λ = 560 nm using an automatic Modulus™ Microplate Photometer (Turner BioSystems, CA, USA). The data were analyzed by calculating the percentage of MTT reduction compared with the control (untreated cells, 100% viability).

Determination of proinflammatory cytokines
To evaluate the anti-inflammatory activity of the derivatized NPs (TAT-NPs, pAntp-NPs and G2-NPs) and the commercial drug (Isoptoflucon), HCE-2 cells were seeded (1 × 10 5 cell·ml −1 ) in 12-well plates and grown until 90% confluency. The different NPs were added to the culture medium at 50 μg·ml −1 (drug concentration loaded NPs) and inflammation was induced with LPS (1 μg·ml −1 ). Cells stimulated only with LPS were considered the positive control and untreated cells as the negative control. After 48 h incubation, the supernatants were collected and centrifuged (10,000 r.p.m for 5 min) at 4 • C and stored at 80 • C until use. Secreted levels of the pro-inflammatory cytokines IL-1β, IL-6, IL-8 and TNF-α were quantified by ELISA sets (BD Biosciences, CA, USA) according to manufacturer's instructions. The results were expressed as pg·ml −1 .

Cellular uptake study
To assess internalization of NPs in HCE-2 cells, 1 × 10 5 cell·ml −1 HCE-2 were grown in eight-well chamber slider (ibidi R , Gräfelfing, Germany) until approximately 80% confluence and then incubated with 50 μg·ml −1 (drug concentration loaded NPs) Rho NPs ( Rho TAT-NPs, Rho pAntp-NPs, Rho G2-NPs and Rho FMT-m-PEG-PLGA-NPs) at 37 • C for 48 h. Noninternalized Rho NPs were then removed by three PBS washes and cells were fixed with 3% paraformaldehyde for 30 min at 25 • C. Afterwards, cells were washed three times with PBS and the nuclei were stained with 4',6-diamidino-2-phenylindole (DAPI) for 15 min at 25 • C. Finally, the cells were washed three times and mounting solution (PBS) was added for microscopic analysis. Images were acquired using a Leica TCS SP5 confocal laser scanning microscopy (Leica Microsystems, Wetzlar, Germany) with a 63x oil immersion objective lens.
In vivo study: ocular uptake C57BL/6J wild-type male mice of 3 months of age were used. Animals were under controlled temperature, humidity and light conditions with ad lib access to food and water. Mice were treated in accordance to ethical guidelines of the European Community Council Directive 86/609/EEC and the procedures established by the Department of Agriculture, Branch and Fisheries of the Government of Catalonia (approved by the ethics committee of University of Barcelona).
Fluorescent formulations (50 μl) were administered in the right eye and the contralateral eye was used as a control. After 48 h of treatment, all animals were sacrificed. The eyes were enucleated and fixed in 4% paraformaldehyde for 24 h at 4 • C. Then, the samples were embedded in optimal cutting temperature compound and stored at −40 • C. Horizontal sections of 20 mm were collected using a Leica CM3050s cryostat (Leica Microsystems, Wetzlar, Germany). The bright and fluorescent fields of the sections were taken by a fluorescent inverted microscope (Leica DMI4000 B).

Statistical analysis
Multiple comparisons were developed using two-way analysis of variance (ANOVA) with Tukey post hoc test with a significance of p < 0.05 after having confirmed the normality and equality of variances by Bartlett in the groups. All analyzed data were presented as mean ± SD. GraphPad Prism 6.01 software and ImageJ was used to analyze the data and images, respectively.

Synthesis of Rho-PLGA & m-PEG-PLGA
The yield of the synthesis of the polymers Rho-PLGA and m-PEG-PLGA after lyophilization was 85.7-92.1% and 81.4-90.1%, respectively. The H-NMR analysis identified the characteristic peaks of each polymer synthesized. Particularly, in the PLGA RG 503H the peaks are visualized at 5.21 ppm for CH proton of lactide, 4.90 ppm for CH 2 proton of glycolide and 1.47 ppm for CH 3 proton of lactide (Supplementary Figure 1A). In the case of the NHS-PLGA, the same previous peaks are revealed plus a peak at 2.70 ppm corresponding to the CH 2 proton of the NHS (Supplementary Figure 1B). The polymer synthesized m-PEG-PLGA (Figure 1), showed the same peaks as the PLGA RG 503H with the presence of the characteristics of m-PEG (Supplementary Figure 1C) Figure 1D) to the NHS-PLGA polymer is confirmed by the absence of the CH 2 proton NHS peak at 2.70 ppm. The PEGylation efficiency of the NHS-PLGA polymer was between 26.9 and 30.1% (MR between 7.6 and 9.0%).

Structural studies of polymers
To investigate the degree of crystallization after the conjugation between the polymers NHS-PLGA and m-PEG, a study was carried out in XRD. In Figure 2A,

Characterization of FMT-m-PEG-PLGA-NPs
Once the FMT-m-PEG-PLGA-NPs and Rho NPs were synthesized and resuspended in PBS, their physicochemical characteristics such as Zav, PI, ZP and EE were evaluated (Table 1).

Synthesis of CPPs
Amidated CPPs that possess in their sequence 4 glycine as a spacer and a cysteine in the N-terminus to be linked with the maleimide were synthesized by the solid-phase method with an automatic synthesizer. The molecular weight of the CPPs with a purity of 90% was analyzed by MALDI-TOF/TOF (Supplementary Figure 2).  Table 1.

Cytotoxicity study
Toxicity of NPs on the HCE-2 cell line was evaluated at two different concentrations using the MTT viability assay. Cell viability was analyzed at 24 and 48 h ( Figure 3). Results showed that at 24 h, none of the NPs conjugated with CPPs was cytotoxic as cell viability was kept close to 100% when assayed both at low (5 μg·ml −1 ) and high (50 μg·ml −1 ) FMT concentration. At 48 h, cell viability was reduced to about 80-90% with TAT-NPs and pAntp-NPs at 5 μg·ml −1 and close to 70% at 50 μg·ml −1 . In contrast, no toxicity was observed for G2-NPs. In this case, the increased MTT activity compared with untreated cells could be attributed to a proliferative effect of theses NPs. Cytotoxicity of Isoptoflucon eye drops (containing FMT as the active drug) was also analyzed in parallel at the same concentrations, and cell viability was of around 82% (at 5 μg·ml −1 ) and 38% (at 50 μg·ml −1 concentration) at 24 h, and 72 and 28%, respectively, when cells were incubated for 48 h. These results indicate that encapsulation of FMT in NPs reduces cytotoxicity of the commercial drug, as it avoids the use of preservatives such as benzalkonium chloride or solubilizers such as polysorbate 80, which are excipients that decrease cell viability [28,29].

Evaluation of the anti-inflammatory activity of NPs
Since LPS induces inflammation via TLR4 receptors, we sought to evaluate the capability of NPs conjugated with CPPs to inhibit the inflammatory response by analysing secreted IL-1β, IL-6 , IL-8 and TNF-α cytokines in LPS-stimulated HCE-2 cells. Results are shown in Figure 4. In the absence of NPs (positive control) LPS induced a significant increase in all cytokines analyzed. The TAT-NPs, pAntp-NPs and G2-NPs significantly decreased the expression of IL-1β similarly to Isoptoflucon (p < 0.001) ( Figure 4A). Concerning IL-6, IL-8 and TNF-α, secreted levels were reduced by TAT-NPs and G2-NPs, whereas pAntp-NPs only reduced IL-8 although to a lesser extent than the other NPs ( Figure 4B-D). In the case of TNF-α ( Figure 4D), the formulation pAntp-NPs showed a greater cytokine expression than the positive control (LPS-treated cells). Overall, these results indicated that the best anti-inflammatory effect was achieved with G2-NPs. In this case, the reduction of LPS-induced cytokine levels was close to that observed for Isoptoflucon, except for IL-8. The formulations developed are in accordance with other nanostructured systems such as nanosponges, which improve the activity of the encapsulated compound in contrast to the free compound [30].

Discussion
In the present study, the polymers Rho-PLGA and m-PEG-PLGA were successfully synthesized with high performance using the carbodiimide pathway as a conjugation strategy [31,32]. The conjugation of both Rho and m-PEG to NHS-PLGA was confirmed by the removal of the CH 2 group from the NHS in the H-NMR spectrum (Figure 1 & Supplementary Figure 1D). Furthermore, the m-PEG moiety was identified in the PLGA due to the appearance of ethylene glycol and maleimide peaks (Figure 1), with a PEGylation degree close to 30% (MR ∼9%) [33,34]. The XRD and FTIR profiles confirmed the information obtained by H-NMR regarding the synthesis of the m-PEG- PLGA polymer. The diffractogram of the m-PEG-PLGA ( Figure 2A) showed a profile of amorphous characteristics without the presence of crystalline peaks of PEG, agreeing with other authors that the presence of less than 10% (m/m) of PEG in the polymer and its peaks are not visualized in the XRD or in the FTIR spectrum [35][36][37]. Despite this fact, the structural change of the NHS-PLGA to the m-PEG-PLGA was evidenced in the FTIR and H-NMR analysis by the elimination of C=O and C-O-N peaks of the NHS, confirming the union m-PEG to the polymer ( Figure 2B). The developed FMT-m-PEG-PLGA-NPs displayed a Zav less than 150 nm with a monodisperse population distribution, a negative surface charge and EE close to 99% (0.5 mg·ml −1 ). It has been described that particles with sizes greater than 10 μm could cause ocular irritation. Hence, the NPs elaborated in this study would circumvent such an effect. Moreover, the agglomeration and/or precipitation of NPs would be avoided with their negative ZP [38,39]. The CPPs TAT 49-57 , pAntp [43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58] and G2 characterized by MALDI-TOF/TOF presented an m/z value of 1613.296, 2520.192 and 1983.102, respectively (Supplementary Figure 2). These CPPs, properly purified and characterized, were conjugated with the FMT-m-PEG-PLGA-NPs. The monodisperses system was maintained in the final NPs and the Zav increased by an average of 17 nm more. Furthermore, the conjugation of the CPPs to the NPs was evidenced both by the change of the Zav and by the change of the ZP to a less negative value by the cationic characteristics of the CPPs (Table 1) [17]. Through H-NMR analysis, it was possible to calculate the MR of the functionalized CPPs in the NPs. The pAntp [43][44][45][46][47][48][49][50][51][52][53][54][55][56][57][58] (MR = 4.6%) was the CPP that was conjugated to a greater degree in mass followed by TAT 49-57 (MR = 2.0%) and G2 (MR = 0.6%). These data indicate that the maleimide-cysteine conjugation pathway is a thermodynamically favorable reaction that allows the conjugation of small amounts of peptides to polymeric NPs [40][41][42].
Functionalized NPs with CPPs did not show cytotoxicity at both low and high concentration (5 and 50 μg·ml −1 ) at 24 h (Figure 3). At the same incubation time, the commercial drug tested at the low concentration reduced cell viability close to 84%, a value that was significantly different (p < 0.05) to the formulation pAntp-NPs. At a higher concentration, Isoptoflucon increased its cytotoxicity. This may be due to the composition of the commercial drug, which contains excipients such as benzalkonium chloride and polysorbate 80 that decrease cell viability [28,29,43]. After 48 h treatment, cells incubated with 5 μg·ml −1 of the formulations TAT-NPs, pAntp-NPs and Isoptoflucon displayed viability rates close to 80%. At 50 μg·ml −1 , the cytotoxicity of the commercial drug ( Figure 3B) was more evident than that of the formulations TAT-NPs, pAntp-NPs and G2-NPs (p < 0.0001). The G2-NPs formulation had no cytotoxic effects as cell viability was kept around 100% at both low and high concentrations. At 48 h viability values were even higher than 100%, which can particularly be due to the nonconjugated sites of the maleimide with G2. The low percentage of MR of G2 in the polymer m-PEG-PLGA in comparison to the other CPPs indicate that the accessible sites of unreacted maleimide may cause an increase in the cellular metabolism and therefore a greater MTT response [44]. Supplementary Figure 4 shows that the formulation without conjugated CPP (FMT-m-PEG-PLGA-NPs) with all the sites of unconjugated maleimide had greater MTT values than G2 (p < 0.01) and the other formulations (p < 0.0001) when assayed in HCE-2 cells at a concentration of 50 μg·ml −1 for 48 h.
Overall, this study shows that G2-NPs is the formulation that has a greater effect in the reduction of the proinflammatory cytokines tested, followed by TAT-NPs. The G2 peptide, although not described as a CPP, has been characterized as an antiviral peptide and contain the conserved features of a CPP, such as a 5-30 amino acid sequence rich in Arg residues and positively charged (cationic) [22,46]. This study provides evidence that G2 and TAT allow the internalization of NPs in HCE-2 cells through the creation of transmembrane pores, thus increasing the anti-inflammatory efficacy. In fact, in vitro/in vivo uptake studies confirmed that Rho TAT-NPs and Rho G2-NPs were the formulations that displayed higher internalization scores. In vitro studies in HCE-2 cells revealed that the highest density of internalized NPs, represented by the intensity of emitted fluorescence, was achieved with Rho TAT-NPs ( Figure 5). The other formulations were internalized but with a lower extent, without fluorescence accumulation as in the case of Rho TAT-NPs and Rho G2-NPs. At the in vivo level (Figure 6), fluorescence intensity of Rho G2-NPs increased with respect to the in vitro study. This effect could be mainly due to unreacted maleimide (without G2), which can be conjugated with the sulfhydryl groups of mucin present in the lacrimal film [47]. This would increase the residence time in the cornea, helping internalization of the NPs in the ocular internal tissues. Analysis of the nonfunctionalized NPs ( Rho FMT-m-PEG-PLGA-NPs) also reflects an increase in the internalized NPs signal, that could also be attributed to their interaction with mucin. At last, a certain number of pAntp-NPs manages to reach the posterior segment of the eye (optic nerve) without being retained at the cornea level.

Future perspective
In this study, the efficacy of using CPPs conjugated to polymeric NPs to increase the cellular uptake into the posterior segment of the eye has been demonstrated. This research has showed that the presence of G2 linked to FMT-loaded NPs increased the ocular penetration thanks to a further link with mucin glycoproteins. A future investigation, we suggest, be related to explore the physicochemical characteristics of G2-nanostructures system as new noninvasive drug delivery vehicle. Specifically, to study whether the antimicrobial properties of G2 are preserved, increasing the synergistic effect of antimicrobial drugs both at the pharmacological level and increasing their cellular internalization.

Summary points
• Cell-penetrating peptides linked to polymeric nanoparticles increase the cellular uptake.
• Fluorometholone-loaded polymeric nanoparticles are an effective drug delivery system to improve the treatment of inflammatory eye diseases. • G2, antiviral peptide, conjugated with polymeric nanoparticles, allowed internalization to the posterior segment of the eye. • The functionalization of nanoparticles with cell-penetrating peptides decreased the expression of TNF-α, IL-6, IL-8.
• TAT-NPs and G2-NPs showed an increasing of fluorescence signal in the human corneal epithelial cells as result of major intercellular accumulation. • The developed fluorometholone-loaded PEG-PLGA nanoparticles displayed a monodisperse distribution, negative charge and entrapment efficiency close to 99%. • High performance was obtained using the carbodiimide pathway for the synthesis of rhodamine-PLGA and maleimide-PEG-PLGA polymers. • In vivo and in vitro uptake studies showed that TAT-NPs and G2-NPs were formulations with the highest density of internalized NPs.

Supplementary data
To view the supplementary data that accompany this paper please visit the journal website at: www.futuremedicine.com/doi/sup No writing assistance was utilized in the production of this manuscript.

Ethical conduct of research
The animal work involved in the in vivo ocular uptake study gained the approved by the ethics committee of the University of