Nanosized Dendritic Polyguanidilyated Translocators for Enhanced Solubility, Permeability, and Delivery of Gatifloxacin

Abstract

Purpose.: Dendrimeric polyguanidilyated translocators (DPTs) are nanosized novel dendrimers that efficiently translocate molecules across biological barriers. The purpose of this study was to develop a DPT that could serve as an ophthalmic delivery vehicle for gatifloxacin and to evaluate its in vitro and in vivo delivery after topical application.

Methods.: The gatifloxacin (GFX) solubility-enhancing property of a six-guanidine group–containing dendrimer (g6 DPT) was investigated as a function of pH and dendrimer concentration. Mechanisms of drug interaction with the dendrimer were investigated by using isothermal titration calorimetry (ITC), Fourier-transformed infrared spectroscopy (FTIR), and nuclear magnetic resonance spectroscopy (NMR). Permeability of the dendrimer was assessed in human corneal epithelial cells (HCECs) and across isolated bovine sclera-choroid-RPE (SCRPE). In vitro efficacy of the dendrimer formulation was evaluated with a time-to-kill assay for methicillin resistant Staphylococcus aureus (MRSA). In vivo delivery of GFX in a dendrimer eye drop formulation was studied in New Zealand White rabbits after a single dose or multiple doses over 3 weeks. Drug levels in various ocular tissues were analyzed by liquid chromatography–tandem mass spectrometry (LC-MS/MS).

Results.: An optimized DPT-GFX formulation (final pH 5.9, no preservative) increased GFX solubility by fourfold. The dendrimer formed isotonically stable, nanosized (346-nm) complexes with GFX via ionic bond, hydrogen bond, and hydrophobic interactions. The dendrimer gained rapid entry into the HCECs (within 5 minutes) and increased the transport of GFX by 40% across the SCRPE in 6 hours. DPT-GFX exhibited a three times faster killing rate for MRSA when compared with GFX alone. In vivo administration of DPT-GFX (1.2% wt/vol) resulted in ∼13-fold, and ∼2-fold higher areas under the curve (AUCs) for tissue concentrations in conjunctiva and cornea, respectively, when compared with GFX (0.3%) after a single dose. Further, a single dose of DPT-GFX sustained aqueous humor and vitreous humor drug levels during the 24-hour study, with a t _1/2 of 9 and 32 hours, respectively. After multiple doses, similar advantages were seen with DPT-GFX.

Conclusions.: The DPT forms stable complexes with GFX and enhances its solubility, permeability, anti-MRSA activity, and in vivo delivery, potentially allowing a once-daily dose regimen.

Despite the ready access of the eye surface to drops, delivery of a therapeutic quantity of drugs to the target tissues in the eye is challenging because of the presence of various ocular barriers. ¹ Typically, <5% of the topically administered dose is bioavailable primarily to the eye tissues in the anterior segment, with little or no drug reaching the back of the eye, ² because of the rapid drainage and clearance of a drop from the eye surface and the poor permeability of the ocular surface tissues. Further, low-molecular-weight drugs in their molecular form are cleared rapidly from eye tissues such as the aqueous and vitreous humors. ^3,4 Thus, multiple eye drop doses are typically needed each day to maintain therapeutic concentration at the target site for treating disorders of the anterior segment. For treating vision-threatening posterior segment disorders such as age-related macular degeneration and diabetic retinopathy, repeated injections in the vitreous space ^1,5 or periocular space ^6,7 are increasingly used. Thus, for key therapies targeted to the eye, reducing the frequency of the doses or developing noninvasive approaches are unmet needs. 

Gatifloxacin (GFX), a fourth-generation fluoroquinolone antibiotic, is approved for treatment of bacterial conjunctivitis caused by aerobic Gram-positive and -negative bacteria. ⁸ For effective therapy and to overcome resistance, the recommended dose regimen for Zymar (0.3% gatifloxacin; Allergan, Irvine, CA) eye drops is administration of 1 drop up to eight times daily on days 1 and 2 followed by four times on days 3 through 7 (http://www.allergan.com/assets/pdf/Zymar_pi.pdf). Thus, development of GFX eye drops with a reduced dose frequency, while maintaining or enhancing its antibacterial properties, is an unmet medical need. 

Recent advances in material science are enabling the development of novel drug delivery systems. Dendrimers belong to a novel class of repeatedly branched, treelike structures with a nanosized monodispersity and abundance of surface functional groups. Different classes of dendrimers have been exploited for targeting various types of cancers, ⁹ rheumatoid arthritis, ^10,11 and other serious disorders. ¹² As illustrated in the present study, the structural uniqueness of dendrimers, with units branching from a core with void spaces within and between the branches and an abundance of surface functional groups, make them useful for complexing or entrapping various therapeutic agents. ¹³ Further, the functional groups on the dendrimer's surface can be covalently conjugated to therapeutic agents, imaging or diagnostic agents, and targeting ligands. ¹⁴ The purpose of this study was to develop a novel dendrimer-based eye drop formulation of GFX that allows reduced dose frequency and enhances therapeutic benefits. 

Dendrimers for ocular drug delivery have been explored in a few prior studies. For instance, poly(amidoamine) (PAMAM) dendrimers with carboxylic and hydroxyl surface groups were shown to have longer corneal residence time and increased activity of incorporated drug after topical instillation in albino rabbits. ¹⁵ Lipid-lysine dendrimers complexed with sense oligonucleotide (ODN-1) exhibited significant inhibition of choroidal neovascularization in a rat model after intravitreal injection. ¹⁶ Also, we have recently shown that a single subconjunctival injection of PAMAM-carboplatin nanoparticles was effective in treating transgenic murine retinoblastoma. ¹⁷ Dendrimeric polyguanidilyated translocators (DPTs) are a novel class of dendrimers with triolyl branches and surface guanidine groups. ¹⁸ Three classes of DPTs—g6, g9, and g12, with g representing the surface guanidine groups, not the generation, as conventionally used in dendrimer terminology—have shown similar efficient uptake in HeLa S3 cells, indicating that six guanidine groups (g6) is optimal for enhanced translocation into cells. ¹⁹ Further, DPTs deliver a small molecule (fluorescein), as well as a macromolecule (green fluorescent protein mutant, GFPcys) across barriers in HeLa S3 cells, and exhibit uptake levels similar to GFP-Tat peptide. ¹⁹ Therefore, we used the g6 DPT as a vehicle for GFX to enhance its solubility, corneal and scleral permeability, anti-MRSA (methicillin resistant Staphylococcus aureus) activity, and in vivo delivery. 

Methods

Materials

Gatifloxacin sesquihydrate (molecular weight [MW], 384.47, 98% purity) was purchased from AK Scientific, Inc. (Mountain View, CA). The g6 DPT or DPT henceforth (base MW, 1650.88, 88% purity) was obtained from BioBlocks, Inc. (San Diego, CA). Trifluoroacetic acid was purchased from Sigma-Aldrich (St. Louis, MO). All other solvents and reagents used were of HPLC-grade purity. 

Analytical Method for Simultaneous Analysis of g6 Dendrimer and GFX

A reversed-phase HPLC method was developed for the simultaneous analysis of the g6 dendrimer and GFX. The system included a solvent delivery pump (TM 616; Waters, Milford, MA), a controller (model 600 S; Waters), an autoinjector (model 717 plus; Waters), and a photodiode array detector (PDA; model 996; Waters). The mobile phase used was a gradient mixture of 0.1% trifluoroacetic acid (mobile phase A) and acetonitrile (mobile phase B) pumped at a flow rate of 0.8 mL/min. Analytes were eluted on a C-18 column (250 × 4.6 mm; Microsorb MV 100-5; Varian, Inc., Palo Alto, CA). GFX was detected at 291 nm and g6 dendrimer at 240 nm. 

Preparation of Isotonic Sorensen's Phosphate Buffers

For formulation of a topical solution of the DPT-GFX complex (dendriplex), isotonic Sorensen's phosphate buffer of various pH levels was prepared, as per the United States Pharmacopeia (USP) procedure. Stock solutions of 0.0667 M monobasic sodium phosphate (solution A) and 0.0667 M dibasic sodium phosphate (solution B) were prepared in deionized water. Formulation vehicles with the following pH levels were prepared by mixing different proportions of the above-mentioned solutions: pH 5.65 buffer (90 mL solution A + 10 mL solution B + 0.52 g sodium chloride), pH 6.30 buffer (70 mL solution A + 30 mL solution B + 0.5 g sodium chloride), and pH 7.4 buffer (20 mL solution A + 80 mL solution B + 0.44 g sodium chloride). 

Solubility Determination of GFX in the Presence of the g6 Dendrimer and Particle Size Analysis

The solubility of GFX at increasing g6 dendrimer concentrations was studied according to the method of Higuchi and Connors ²⁰ as described earlier. Briefly, an excess of drug (15–20 mg) was added to 2 mL of phosphate buffer (pH 5.65, 6.30, or 7.37) containing increasing concentrations of g6 dendrimer. Buffers without dendrimer were included as the control, and all experiments were performed in triplicate. The mixture was incubated at 25°C with rotation at 100 rpm until equilibrium was reached (48 hours). The solution was filtered through a 0.45-μm filter, and the filtrate was analyzed for drug content by HPLC. The final pH levels of the solubility samples were recorded at the end of the experiment and the particle size of the final formulation was measured (Zetasizer Nano ZS Particle Size Analyzer; Malvern Instruments, Ltd., Worcestershire, UK). 

Binding Characterization of GFX to the g6 Dendrimer by Isothermal Titration Calorimetry (ITC)

Titrations were performed using an isothermal titration calorimeter (VP-ITC; MicroCal, Piscataway, NJ). All the samples (g6 dendrimer and drug solution in pH 6.3 buffer, pH 6.3 buffer alone, and deionized water) were degassed by stirring them in a vacuum before use. The reference cell was loaded with water and the sample cell was loaded with g6 dendrimer solution in pH 6.3 buffer (1 mg/mL). The syringe (autopipette) was filled with GFX solution in pH 6.3 buffer (2 mg/mL). The heat of the reaction was obtained by making 19 consecutive injections (15 μL each) of the drug in buffer solution at 240-second intervals into the sample cell that contained the dendrimer. The injection syringe (with a paddle mounted) stirred the solutions at 300 rpm, to ensure immediate mixing, and the experiment was performed at 30°C. The thermal peaks after the titration were integrated by software (Origin 7.0; OriginLab, Northampton, MA) supplied with the instrument. Dilution experiments were performed by identical injections of drug solution into the sample cell that contained buffer (pH 6.3) alone and was subtracted from the drug–dendrimer data. For determining the nature of drug–dendrimer interactions, we performed the titration experiments in the presence of 10 mM NaCl (to inhibit ionic interactions) or 1 mM Triton X-100 (to inhibit hydrophobic interactions). 

Determination of the Mechanism of Interaction by Fourier-Transformed Infrared (FTIR) and Proton Nuclear Magnetic Resonance Spectroscopy (¹H-NMR)

Infrared spectra were acquired by using a Fourier-transform infrared spectrometer (MB-series; ABB Bomem, Quebec City, QB, Canada). Lyophilized solubility assay samples including drug–dendrimer complex (dendriplex or DPT-GFX), g6 dendrimer, and drug samples were analyzed in KBr pellets. A 128-scan interferogram was collected for each spectrum in single-beam mode with a resolution of 4 cm⁻¹. Background spectrum was subtracted from each sample spectrum, followed by data analysis. The proton NMR spectra of samples were recorded with a 500-MHz spectrometer (Inova; Varian, Inc.). GFX, g6 dendrimer, or GFX+g6 dendrimer were dissolved in deuterium oxide (D₂O), and the spectra were recorded. 

In Vitro Release of GFX from Dendriplex in Simulated Tear Fluid

Simulated tear fluid for in vitro release was prepared according to a previous report. ²¹ For in vitro release studies, 0.5 mL of drug solution or DPT-GFX solution in buffer was placed in a dialysis bag (2000 MWCO) and sealed at both ends. The dialysis bag was dropped in 10 mL of simulated tear fluid and incubated at 37°C with stirring at 150 rpm. At different time intervals, 1 mL of sample was taken and replaced with fresh buffer maintained at the same temperature. The experiment was performed in triplicate. The GFX released was analyzed by HPLC. 

Permeability of GFX and g6 Dendrimer across SCRPE

Freshly excised bovine eyeballs were purchased from G&C Packing Co. (Colorado Springs, CO), and the experiments were initiated within 3 to 4 hours of death. Adherent tissues were carefully removed, and the SCRPE tissue was isolated and mounted in Ussing chambers with the scleral side facing the donor compartment and the RPE layer facing the receiver compartment. Exactly 1.5 mL of GFX or DPT-GFX was added to the donor chamber and the same volume of blank assay buffer was added to the receiver chamber at the same time. The surface area of the tissue exposed to the sample solution was 0.64 cm². A 200-μL aliquot of sample was collected after 1, 2, 3, 4, and 6 hours from the receiver chamber and replaced immediately with the same volume of fresh assay buffer. After the 6-hour time point, entire samples from donor chambers were also collected. At the end of the transport study, the tissues were carefully collected and stored at −20°C until analysis by HPLC. For extraction of the drug and the dendrimer from the tissues, whole tissue was homogenized in 2 mL of pH 6.3 buffer and sonicated in a water bath for 30 minutes. To the homogenate was added 1 mL of acetonitrile, followed by vortexing for 30 minutes. The content was centrifuged at 12,000 rpm for 30 minutes, and the clear supernatant was collected and injected onto a HPLC column for analysis. 

Confocal Microscopy of FITC-g6 DPT Uptake in HCECs

Human corneal epithelial cells (HCECs) (a gift from Vasilis Vasiliou, University of Colorado Denver) were plated on coverslips in 12-well plates 12 hours before treatment. The cells were treated with a mixture of 4.31 mM of g6 dendrimer and 0.01 mM of FITC-g6 DPT for 5, 30, and 60 minutes at 37°C, followed by washing with 1 mL of PBS four times. The cells were fixed with 4% paraformaldehyde for 15 minutes, counterstained with DAPI at room temperature, and extensively washed with PBS. Finally, the coverslips were mounted on slides (FluorPreserve; Calbiochem, San Diego, CA) and allowed to dry in the dark. Confocal images were obtained with a confocal microscope (D Eclipse C1; Nikon, Tokyo, Japan). 

Antibacterial Activity by Time-to-Kill Assay

The bactericidal activity of DPT-GFX was estimated by using the time-to-kill assay. ²² Initially, the minimum inhibitory concentration (MIC) of the GFX, g6 dendrimer, and DPT-GFX were determined by the broth microdilution method. For antibacterial assay, approximately 5 × 10⁵ to 5 × 10⁶ CFU/mL bacteria were incubated in the presence of GFX and DPT-GFX at 4× MIC (minimum inhibitory concentration) as determined by broth microdilution testing. g6 DPT alone was included at a constant concentration (1024 μg/mL) as the control. Viable bacteria were quantitated at 0, 5, 15, 30, 45, 60, 120, and 360 minutes after inoculation by serial dilution plating. The rate (time to kill) and the extent of killing were determined from the plot of viable counts (log₁₀ CFU/mL) against time (minutes). Bactericidal activity was defined as a ≥3-log₁₀ decrease in colony-forming units per milliliter relative to the initial inocula. The minimum concentration of drug needed to kill most (>99.9%) of the viable organisms after incubation for a fixed length of time under the given set of conditions was defined as the minimal bactericidal concentration (MBC). 

Acute-Dose In Vivo Pharmacokinetic Study

All animal studies were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Male New Zealand White rabbits in the weight range of 1.8 to 3 kg were used for the acute dose study. Animals were divided into six groups (three animals each). Left eyes of rabbits were dosed with two 50-μL drops of DPT-GFX (1.2% wt/vol GFX) at 5-minute intervals. The animals were euthanatized at 15 minutes, 30 minutes, and 1, 6, 12, and 24 hours after the last dose. Aqueous humor was collected immediately from both the eyes, followed by enucleation. The enucleated eyes were immediately frozen in dry ice and stored at −80°C before tissue isolation. Frozen eyes were dissected to isolate cornea, conjunctiva, vitreous humor, and sclera. Isolated tissues were stored at −80°C until analysis by LC-MS/MS. 

Multiple-Dose In Vivo Pharmacokinetic Study

Male New Zealand White rabbits in the weight range of 1.8 to 3 kg were used for the multiple-dose study. Animals were divided into six groups (three animals each). One 50-μL drop of Zymar (0.3% wt/vol GFX; Allergan) was instilled into the right eyes of the rabbits, and the left eyes received one 50-μL drop of DPT-GFX (1.2% wt/vol GFX) three times a day for 7 days (total 21 drops). Animals were euthanatized at 15 minutes, 30 minutes, and 1, 6, 12, and 24 hours after the last dose (21st dose). Aqueous humor was collected immediately from both the eyes, followed by enucleation. The enucleated eyes were immediately frozen in dry ice and stored at −80°C before tissue isolation. Frozen eyes were dissected to isolate cornea, conjunctiva, vitreous humor, and sclera. Isolated tissues were stored at −80°C until analysis by LC-MS/MS. 

GFX Extraction from Ocular Tissues

To 50 μL of aqueous humor, 100 μL of vitreous humor, or a weighed quantity of tissue, an aliquot of 12.5 μL of 5 μg/mL moxifloxacin (internal standard) solution was added. The samples were vortexed for 5 minutes at medium speed. To this solution, 1200 μL of ethyl acetate was added and vortexed for 15 minutes in a multitube vortexer followed by centrifugation at 10,000 rpm for 10 minutes. The ethyl acetate layer was separated into a clean test tube and evaporated under nitrogen stream. The residue was reconstituted with 250 μL of acetonitrile, and a 10-μL aliquot of this solution was injected into LC-MS/MS for analysis. 

LC-MS/MS Conditions for GFX Analysis

LC-MS/MS analysis was performed on a system (200 Series; Perkin Elmer, Boston, MA) equipped with two pumps connected to a low-pressure mixer, online vacuum degasser, and 96-well plate auto-sampler. An LC-MS/MS system (API 3000) equipped with electrospray source and operated in positive ionization mode was used as the detector. Commercial system management software (Analyst 1.42; Applied Biosystems, Foster City, CA) was used for instrument control, data acquisition, and processing. The chromatographic separations of analytes were performed on a C5 column (Supelco; 100 × 2.1 mm; Sigma-Aldrich). The mobile phase consisted of water (mobile phase A) and acetonitrile (mobile phase B). A linear gradient elution with a total run time of 4 minutes was set as follows: 60% of A for 1 minute; linear to 20% by 1.2 minutes; and 20% for the next 1 minute, followed by an increase to 60% in 0.3 minute with 1.5 minutes of re-equilibration time before the next injection. The mobile phase flow rate was 0.3 mL/min. The autosampler was maintained at room temperature and 10 μL of sample was injected. The following MRM (multiple reaction monitoring) transitions were used for monitoring GFX (376/358) and moxifloxacin (402/384). The retention time of GFX was ∼1.42 minutes, and that of the internal standard was ∼1.57 minutes. Standard curves were constructed by spiking known quantity of drug solution in the respective tissue matrix. Linearity was observed in the range of 10 and 2000 ng/mL for various tissues. 

Pharmacokinetic and Statistical Analysis

Pharmacokinetic analysis was performed by noncompartmental analysis (WinNonlin software, ver. 1.5; Scientific Consulting, Inc., Gaithersburg, MD). Groups were compared for statistical significance by Student's t-test (α = 0.05). 

Results

Simultaneous Analysis of GFX and the g6 Dendrimer by HPLC

Since the g6 dendrimer was highly soluble in water, and GFX was sparingly soluble, a gradient elution was used for the simultaneous analysis of both the g6 dendrimer and GFX. The gradient conditions used were 13% of phase A for 3 minutes; gradient to 50% in 3.5 minutes; maintenance at 50% for the next 3 minutes (until 6.5 minutes); and a decrease to 13% in 0.5 minute with 8 minutes' re-equilibration time before the next injection. The HPLC method provided simultaneous analysis of GFX and the g6 dendrimer. Under these conditions, GFX eluted at 4.200 ± 0.107 minutes and the g6 dendrimer eluted at 9.869 ± 0.117 minutes. No other interfering peaks were observed in the retention times for GFX and the g6 dendrimer (data not shown). The coefficients of variation of the drug and g6 dendrimer were within ±3%, indicating the reproducibility of the HPLC method. The signal linearity for GFX was observed in the range of 0.36 to 46 μg/mL, and the g6 dendrimer linearity was in the range of 1.56 to 100 μg/mL. The intraday assay was performed by injecting two sets of the linearity standards on the same day. The interday assay was performed by injecting the linearity standards on the next day, and the results were compared with those of the intraday samples. The intra- and interday precision of GFX was within 6%. The intra- and interday precision of g6 dendrimer was <16%, except at the lower limit of quantitation (LLOQ). This HPLC method was used for the simultaneous determination and estimation of GFX and g6 dendrimer from solubility and transport study samples. 

Solubility Determination of GFX in the Presence of g6 Dendrimer

The g6 dendrimer was freely soluble in the phosphate buffers of all three pH values (>200 mg/mL). All solubility study samples were diluted by 100- to 400-fold in a mixture of acetonitrile and water (1:1, vol/vol), so that the diluted concentrations were in the linearity range of standard curve. The results of the analysis were summarized in Figure 1a. Solubility of GFX in the absence of dendrimer was 4.35 ± 0.16 mg/mL in pH 5.65 buffer and decreased with increasing pH values (2.45 ± 0.04 mg/mL in pH 7.37 buffer). The presence of a low amount of g6 dendrimer (1 mg/mL) did not improve the solubility of GFX. However, at a higher dendrimer concentration (10 mg/mL), solubility of GFX was 2.45-fold higher in pH 5.65 buffer and 2.8-fold higher in pH 6.3 buffer. A 1.6-fold increase in GFX solubility was achieved in the pH 7.4 buffer. In light of these results, further investigations were performed with the pH 6.3 buffer at increasing dendrimer concentrations (15 and 20 mg/mL). As shown in Figure 1b, a linear increase (R ² = 0.9989) in GFX solubility was observed in the range of 1 to 20 mg/mL (0.6–12 μM) of g6 dendrimer. The solubility isotherm of DPT-GFX in pH 6.3 buffer exhibited an A_L type curve indicating a linear relationship in GFX solubility with increasing g6 dendrimer levels. 

Figure 1.

 

(a) Solubility of GFX in phosphate buffers in the presence of g6 dendrimer. (b) Solubility isotherm of GFX in the presence of the g6 dendrimer (pH 6.3 buffer) indicating the formation of a higher order complex. (c) Change in pH values in the presence and absence of the g6 dendrimer and GFX.

Alteration in the pH of Buffers after Addition of g6 Dendrimer and GFX and Particle Size Measurement

The effect of addition of GFX and g6 dendrimer on the pH of buffers is shown in Figure 1c. In all three buffers, the pH value decreased after the addition of a higher amount (10 mg/mL) of g6 dendrimer, with the effect being more pronounced with lower-pH buffers. The addition of GFX alone (in the absence of dendrimer) caused a slight increase in the pH of the 5.65 buffer to pH 6.11; however, the pH of other buffers remained unaltered. The pH of solubility samples at 1 mg/mL g6 dendrimer concentration was almost similar to that of the samples with GFX. Drug samples containing 10 mg/mL g6 dendrimer caused a pH decline in the 7.37 buffer to 6.65 and in the 6.3 buffer to 6.02. In vitro transport and release studies were performed with the 10-mg/mL dendrimer solution (pH 6.02), and the in vivo studies were conducted with the 15-mg/mL dendrimer formulation (pH 5.9). The average particle size of the final formulation was 345.9 ± 53.5 nm, with a polydispersity index of 0.552 ± 0.10. 

Binding Characteristics of GFX to the g6 Dendrimer by ITC

Raw data for the titration of GFX with the g6 dendrimer in the pH 6.3 buffer is shown in Figure 2a, along with the integrated peaks. Each peak in the binding isotherm represents a single injection of drug into the dendrimer solution. The binding of GFX to g6 dendrimer exhibited both exothermic and endothermic peaks. A standard nonlinear least-squares regression binding model involving three sequential binding sites fitted well to the data. The binding constants (K in M⁻¹), change in enthalpy (ΔH in cal/mole), and change in entropy (ΔS in cal K⁻¹ mole⁻¹) for all the binding sites are summarized in Table 1. The binding constant (K2) was lower for one site (2249 M⁻¹), when compared with the other two sites. The enthalpy value was negative for one site (ΔH1) and positive for two others (ΔH2 and ΔH3), whereas the entropy values remained positive for all three sites. Negative enthalpy indicates enthalpically favored binding. 

Figure 2.

 

(a) Data from the ITC titration of the g6 dendrimer and GFX in pH 6.3 buffer (top) and the integrated heat profile of the titration (bottom). The same titration was repeated in the presence of (b) 10 mM NaCl, to study the ionic interactions, and (c) 1 mM Triton X-100, to investigate the hydrophobic interactions.

Table 1.

 

Thermodynamic Parameters Associated with Binding of GFX with the g6 Dendrimer

Binding Site	Parameter	pH 6.3 Buffer	pH 6.3 Buffer + NaCl (10 mM)	pH 6.3 Buffer + Triton-X (1 mM)
1	K (M−1)	1.374E5	3.688E5	3.537E5
	ΔH (kcal/M)	−40.92	−1.971	−103.6
	ΔS (cal/K/M)	23.37	25.46	25.04
2	K (M−1)	2249	4.892E4	9948
	ΔH (kcal/M)	10.18	109.4	166
	ΔS (cal/K/M)	15.37	21.82	18.84
3	K (M−1)	1.135E6	2.201E5	5.029E5
	ΔH (kcal/M)	70.49	1.504	−274.6
	ΔS (cal/K/M)	27.94	24.45	25.18

 The binding isotherm was fit to a three sequential binding site models on the basis of the goodness-of-fit metric, χ² distribution test. K, binding constant; ΔH, change in enthalpy; ΔS, change in entropy.

Binding in the presence of excess NaCl, inhibited exothermic peaks (Fig. 2b). Also, a significant shift in enthalpy was observed for the first binding site (ΔH1 = −40.92 to −1.97), and the enthalpies at the other two binding sites (ΔH2 and ΔH3) remained positive. No specific changes were observed in the entropies of all the binding sites with the addition of NaCl. In the presence of 1 mM Triton X-100, only exothermic peaks were observed in the ITC curves (Fig. 2c). The other important feature was the reversal of enthalpy from positive to negative (ΔH3 = −274.6) at one of the binding sites, when the titration was performed in the presence of Triton X-100. 

FTIR Characterization of the GFX–g6 Dendrimer Complex

The FTIR spectra of the g6 dendrimer, GFX, and DPT-GFX are shown in Figure 3a. The g6 dendrimer (structure shown in Fig. 4a) has surface guanidine groups and amide linkages. The bands between 3100 and 3500 cm⁻¹ (3259 and 3154 cm⁻¹), seen in the FTIR spectrum of the g6 dendrimer, are due to the NH stretch arising from the guanidine groups and amide bonds. The bands at 1658 and 1439 cm⁻¹ characteristic of CN stretch and CN stretch, respectively, confirm the presence of guanidine groups. Other characteristic amide bands are evident at 1649 cm⁻¹ (CO stretch) and at 1549 cm⁻¹ (NH bending). The band observed at 2878 cm⁻¹ is due to the aliphatic CH stretch of the repeating CH₂ groups in the branching units of g6 dendrimer. Prominent features in the FTIR spectrum of GFX (structure shown in Fig. 4b) include the characteristic carboxylic acid bands at 3396 (OH stretch), 1729 (CO stretch), and 1280 cm⁻¹ (CO stretch). The presence of fluoride in GFX can be inferred from the bands between 960 and 1250 cm⁻¹. The aromatic bands (CC stretch) of GFX are evident between 1450 and 1600 cm⁻¹. 

Figure 3.

 

(a) In the FTIR spectrum of DPT-GFX, absence of the OH stretch band at 3396 cm⁻¹ and CO stretch band at 1280 cm⁻¹ (arrows) corresponding to the GFX carboxyl group indicate the involvement of GFX carboxyl group in the formation of the DPT-GFX complex. The NH stretch bands observed at 3259 and 3154 cm⁻¹ in the g6 dendrimer shifted to 3266 and 3166 cm⁻¹, respectively, in the DPT-GFX, indicating involvement of guanidine (NH₂, NH) and amide (NH) groups in the complexing process. (b) ¹H-NMR spectra (D₂O solvent) of g6 dendrimer, GFX, and DPT-GFX. The downfield shift in the protons of (C[b]H₂)₂ cyclopropyl group at 1.00 ppm and the aromatic quinolone ring of GFX 7.805 and 8.588 ppm in the DPT-GFX indicate the entrapment of GFX in the void space between the branches and core of the g6 dendrimer. Also the downfield shift in the OCH₂C[b]H₂, C[b]H₂CO protons of the dendrimer branches at 2.3 to 2.5 ppm further supports the involvement of these groups in forming the hydrogen bonds with GFX.

Figure 4.

 

Structure of (a) g6 dendrimer, (b) GFX, and (c) proposed sites for interactions in the g6 dendrimer.

The FTIR spectrum of the DPT-GFX showed the characteristic bands of both g6 dendrimer and GFX. A distinctive feature in the spectrum of DPT-GFX was the absence of a band at 3396 cm⁻¹ that was observed in the GFX spectrum corresponding to OH stretch of the carboxyl group. This, in conjunction with the absence of carboxyl CO stretch band at 1280 cm⁻¹ in the DPT-GFX spectrum, confirms the involvement of carboxyl group in the formation of GFX complex with g6 dendrimer. Another feature in the spectrum of DPT-GFX is the reduced intensity of bands at 997 and 1061 cm⁻¹ and appearance of a new band at 1170 cm⁻¹ indicating that fluoride group may be involved in the formation of complex with g6 dendrimer. The NH stretch bands observed at 3259 and 3154 cm⁻¹ in the g6 dendrimer shifted to 3266 and 3166 cm⁻¹, respectively, in the DPT-GFX, indicating involvement of guanidine (NH₂, NH) and amide (NH) groups in the complexing process. 

¹H-NMR Characterization of the GFX-g6 Dendrimer Complex

The ¹H-NMR spectra of GFX, g6 dendrimer, and DPT-GFX are shown in Figure 3b. In the GFX spectrum, a proton shift corresponding to (C[b]H₂)₂ of cyclopropyl was observed at 1.00 ppm and that of the methyl C[b]H₃ group was present at 1.174 ppm. The piperazine group protons were evident between 3.3 and 3.7 ppm and the two protons in the quinolone ring of gatifloxacin were indicated by the sharp signals at 7.805 and 8.588 ppm. In the g6 dendrimer spectrum, the methyl C[b]H₃ proton peaks were observed at 1.133 ppm. The protons of C[b]H₂CO and OCH₂C[b]H₂ groups of the dendrimer branches exhibited signal between 2.3 and 2.5 ppm. The aromatic protons in the core of the dendrimer gave signal at 8.359 ppm. The downfield shift in the protons of cyclopropyl group and the aromatic quinolone ring of GFX in the DPT-GFX indicated their proximity to the electronegative atoms (in the core and shell of the dendrimer), which may be due to the entrapment of GFX in the void space between the branches and core of g6 dendrimer. Also the downfield shift in the protons of dendrimer branches (OCH₂C[b]H₂, C[b]H₂CO, and C[b]H₂NHCO) further supports the involvement of these groups in forming the hydrogen bonds with GFX. 

In Vitro Release of GFX from the Dendriplex

The in vitro release profile of GFX from DPT-GFX in simulated tear fluid is shown in Figure 5a. GFX release was significantly lower than that of DPT-GFX when compared with plain solution. At the end of 24 hours, nearly 86% of GFX was released into simulated tear fluid in case of solution, whereas only 67% of GFX was released from DPT-GFX. 

Figure 5.

 

(a) Cumulative release profile of GFX from solution and DPT-GFX in simulated tear fluid. (b) Cumulative % transport of GFX from pH 6.3 solution and DPT-GFX across bovine SCRPE tissues. The receiver side was blank assay buffer at pH 7.4. Each data point represents the mean ± SD (n = 3 or 4).

Permeability of GFX and g6 Dendrimer across Bovine SCRPE

The cumulative transport of GFX in pH 6.3 buffer solution or DPT-GFX across SCRPE tissues is shown in Figure 5b. At the end of 6 hours, 0.68% of GFX was transported across the SCRPE from DPT-GFX, whereas 0.48% of drug was transported from the plain drug solution. No detectable level of g6 dendrimer was observed in the receiver chamber until 4 hours. At the end of 6 hours, ∼0.14% of the g6 dendrimer was transported across the SCRPE. The drug uptake by the SCRPE at the end of 6 hours was 22.63% ± 3.01% from the GFX solution and 17.87% ± 3.70% from DPT-GFX. Also, no quantifiable g6 dendrimer levels were detected in any of these tissues at the end of 6 hours. 

Uptake of FITC-g6 Dendrimer by HCECs

The confocal images of HCECs after incubation with FITC-g6 dendrimer at different time intervals are shown in Figure 6. It is evident from the figure that FITC-g6 DPT gained rapid entry into the HCECs (within 5 minutes; Fig. 6a). Increased intensity in the fluorescence was observed when the incubation time was prolonged to 30 (Fig. 6b) and 60 (Fig. 6c) minutes. 

Figure 6.

 

Confocal images of FITC-g6 DPT uptake in HCECs after incubation at 37°C for (a) 5 minutes, (b) 30 minutes, and (c) 60 minutes followed by washing with 1 mL of PBS four times. The cells were fixed with 4% paraformaldehyde for 15 minutes, counterstained with DAPI at room temperature followed by extensive washing with PBS. Finally, the coverslips were mounted on slides and allowed to dry in the dark before confocal images were obtained.

Antibacterial Study by the Time-to-Kill Method

Results of the time-to-kill assay are summarized in Table 2. The g6 DPT alone did not have any antibacterial activity. Also, the presence of DPT in the complex did not enhance the MIC and MBC of GFX against the bacterial strains tested. A two to four times faster kill time was observed with DPT-GFX against MRSA and Pseudomonas organisms, whereas no difference was observed between GFX and DPT-GFX against Haemophilus influenzae. 

Table 2.

 

Comparative Killing Rates of GFX and DPT-GFX Against Common Organisms

Organism	Strain ID	Phenotype	Specimen Source	MIC (μg/mL)		MBC (μg/mL)		Decreased Time to Kill
				GFX	DPT-GFX	GFX	DPT-GFX
S. aureus	29213			0.06	0.06	0.12	0.12	No difference
S. aureus	2781984	MSSA	Conjunctiva	0.12	0.12	0.12	0.12	No difference
S. aureus	2781985	MRSA, Levo R	Conjunctiva	4	4	4	4	3× faster for DPT-GFX
P. aeruginosa	27853			1	1	2	2	4× faster for DPT-GFX
P. aeruginosa	2781982		Cornea	1	1	1	1	2× faster for DPT-GFX
P. aeruginosa	2781983	Levo R	Conjunctiva	64	128	>128	128	No difference
S. pneumoniae	49619			0.25	0.25	0.25	0.25	1.7× faster for DPT-GFX
S. pneumoniae	2781979	PEN S	Eye	0.25	0.25	0.25	0.25	1.25× faster for DPT-GFX
S. pneumoniae	2781978	PEN R	Eye	0.25	0.25	0.25	0.25	1.5× faster for DPT-GFX
H. influenzae	49247			≤0.015	0.015	≤0.015	≤0.015	No difference
H. influenzae	2781980	b-lac positive	Eye	≤0.015	0.015	≤0.015	0.015	No difference
H. influenzae	8781981	b-lac negative	Eye	≤0.015	0.015	≤0.015	0.015	No difference

 For the time-kill assay, approximately 5 × 10⁵ to 5 × 10⁶ CFU/mL bacteria were incubated in the presence of GFX and DPT-GFX at 4× MIC as determined by broth microdilution testing. The g6 dendrimer alone, included at a constant concentration (1024 μg/mL) as control, did not exert any antibacterial activity. Viable bacteria were quantitated at 0, 5, 15, 30, 45, 60, 120, and 360 minutes after inoculation by serial-dilution plating. Bactericidal activity was defined as a ≥3-log₁₀ CFU/mL decrease relative to the initial inocula. MBC is the minimal concentration of drug needed to kill most (>99.9%) of the viable organisms after incubation for a fixed length of time under the given set of conditions. S. aureus, Staphylococcus aureus; P. aeruginosa, Pseudomonas aeruginosa; S. pneumoniae, Streptococcus pneumoniae; H. influenzae, Haemophilus influenzae.

Acute Dose In Vivo Pharmacokinetic Study

The mean GFX content in various ocular tissues after instillation of two 50-μL drops of DPT-GFX is shown in the Figure 7. High drug levels are evident in the cornea followed by the conjunctiva and sclera. Corneal drug levels were detectable up to 6 hours. In the conjunctiva, detectable levels of GFX were observed until 12 hours. However, the conjunctival drug levels could be obtained only from two animals at the 6- and 12-hour time points. In contrast, scleral drug levels were quantifiable until 1 hour in all three animals. GFX was detected at all observed time intervals (until 24 hours) in both the aqueous and vitreous humors. The aqueous levels were higher than the vitreous levels, but below the tissue levels at all observed time points. The mean pharmacokinetic parameters estimated for DPT-GFX in various ocular tissues are summarized in Table 3. The AUCs were higher for the cornea followed by conjunctiva and aqueous humor in the DPT-GFX group. The half-life of GFX released from DPT-GFX was higher in the vitreous humor (31.6 hours) and aqueous humor (8.9 hours). 

Figure 7.

 

GFX concentration in ocular tissues of rabbits after topical administration of two 50-μL drops of DPT-GFX. Each data point represents the mean ± SD (n = 3).

Table 3.

 

Pharmacokinetic Parameters Estimated for the Acute Dose Study after Topical Administration of Two 50-μL Drops of DPT-GFX (1.2%) to Male New Zealand White Rabbits

PK Parameter	Aqueous Humor	Vitreous Humor	Cornea	Conjunctiva	Sclera
C max, μg/mL or g	1.2004 ± 0.0627	0.1708 ± 0.0379	26.4859 ± 12.0262	10.3419 ± 0.8665	1.6571 ± 0.4263
T max, h	1.00 ± 0.00	0.83 ± 0.2887	0.2500 ± 0.00	0.5833 ± 0.3819	0.4167 ± 0.1443
AUC0-t, μg · h/mL or g	6.7462 ± 0.1927	1.9189 ± 0.4415	60.3349 ± 6.0341	29.5688 ± 23.6959	1.1005 ± 0.2513
t 1/2, h	8.9005 ± 0.0463	31.6111*	1.3877 ± 0.2998	3.4441 ± 3.0440	1.1951 ± 1.2898
AUC0-∞, μg · h/mL or g	8.4183 ± 0.2008	4.4796*	62.2325 ± 5.6931	35.5070 ± 27.1892	3.3089 ± 3.5643
Cl, L or g/h	0.1426 ± 0.0034	0.27*	0.0194 ± 0.0017	0.0855 ± 0.1057	0.6937 ± 0.4615

 Data are expressed as mean pharmacokinetic parameter estimates ± SD; n = 3. Parameters were estimated by using noncompartmental analysis (WinNonlin, version 1.5, Pharsight Inc., Mountain View CA). C _max is the maximum observed drug concentration in a particular tissue; T _max is the time at which C _max occurs; AUC_0-t is the AUC obtained by plotting the concentration-time data where t is the time of last observed drug concentration; Cl is the clearance of drug; AUC_0-∞ is the area under the curve obtained by plotting the concentration-time data extrapolated to time infinity.

*  Drug detected in only two eyes (n = 2).

GFX drug levels in cornea, conjunctiva, and aqueous humor after administration of DPT-GFX (1.2%) in comparison with plain solutions of different strengths (0.1%, 0.3% and 0.5%) of GFX are shown in the Figure 8. GFX solution data from the different strengths were obtained from the NDA application submitted by Allergan (21-493, available at the FDA/CDER [U.S. Food and Drug Administration Center for Drug Evaluation and Research] web site at http://www.fda.gov/cder/). In the NDA study, male Japanese white rabbits were treated topically with a single dose (two 25-μL drops at 5-minute intervals) of ¹⁴C-GFX of various strengths (0.1%, 0.3%, and 0.5%). Ocular tissues—aqueous humor, conjunctiva, and cornea—were collected at 0.25, 0.5, 1, and 2 hours after the dose, followed by radioactivity quantification with a liquid scintillation counter. 

Figure 8.

 

GFX concentration in the (a) cornea, (b) conjunctiva, and (c) aqueous humor of albino rabbits after a single dose (2 drops at 5-minute intervals) of DPT-GFX, in comparison with GFX solution administered at various doses (0.1%, 0.3%, and 0.5%; NDA application 21-493, submitted by Allergan, Irvine, CA). DPT-GFX data were collected up to 24 hours whereas the GFX data (from the NDA application) were reported up to 2 hours. Each data point for DPT-GFX (n = 3) and GFX (n = 4) represents the mean ± SD.

In the present study, GFX was not detectable in the cornea after 2 hours in the case of plain solutions, whereas sustained drug levels were observed until 6 hours after DPT-GFX administration (Fig. 8a). The cases with conjunctiva (Fig. 8b) and aqueous humor were similar (Fig. 8c), where DPT-GFX sustained the drug levels for 12 and 24 hours, respectively. On the other hand, for plain GFX in the Zymar NDA study, drug levels were reported only up to 2 hours in conjunctiva and aqueous humor. The summary of pharmacokinetic parameters estimated in these tissues for the various GFX strengths is provided in Table 4. Note that there was no dose-proportionate increase in the AUC or C _max for the GFX in the absence of DPT. The PK parameters increased when the dose was increased from 0.1% to 0.3%, but a further increase in the dose to 0.5% resulted in a decreased C _max and AUC values. Despite this dose limitation, DPT-GFX with 1.2% wt/vol of GFX, achieved greater drug levels. GFX delivery, as assessed by its AUC_0-∞, was ∼2-fold higher in the cornea and ∼13-fold higher in the conjunctiva when compared with the clinically available strength of 0.3% GFX after a single bilateral dose in albino rabbits (Fig. 8, Table 4). Also, the half-life of GFX was increased by ∼fivefold (3.44 hours) in the conjunctiva after administration of DPT-GFX when compared with 0.3% GFX (0.65 hour), whereas no advantage was observed in the cornea. Similarly, an approximately six- to ninefold increase in drug half-life was noted in the aqueous humor in comparison to 0.1% and 0.5% GFX (half-life could not be determined for 0.3% GFX). 

Table 4.

 

Pharmacokinetic Parameters Estimated for the Acute Dose Study after Topical Administration of Two 50-μL Drops of DPT-GFX (1.2%) in Comparison with the Plain GFX Solution Administered as Two 25-μL Drops at Different Strengths in New Zealand White Rabbits

PK Parameter	GFX Strength (NDA Application)			DPT-GFX 1.2%	DPT-GFX/GFX 0.3%
	0.1%	0.3%	0.5%
Cornea
C max, μg/mL or g	3.81	21.46	12.23	26.49	1.23
T max h	0.25	0.25	0.50	0.25	1
AUC0-t, μg · h/mL or g	4.72	23.04	12.84	60.33	2.62
t 1/2, h	1.03	1.91	0.65	1.39	0.73
MRT0-t, h	0.73	0.54	0.68	0.93	1.72
AUC0-inf, μg · h/mL or g	6.57	35.49	14.97	62.23	1.75
Conjunctiva
C max, μg/mL or g	1.01	4.81	1.74	10.34	2.15
T max, h	0.25	0.25	0.50	0.58	2.32
AUC0-t, μg · h/mL or g	0.75	2.49	1.70	29.57	11.88
t 1/2, h	0.76	0.65	0.47	3.44	5.29
MRT0-t, h	0.70	0.60	0.67	2.46	4.10
AUC0-inf, μg · h/mL or g	0.90	2.74	1.83	35.51	12.96
Aqueous humor
C max, μg/mL or g	0.28	1.12	1.19	1.20	1.07
T max, h	0.50	1.00	0.50	1.00	1
AUC0-t, μg · h/mL or g	0.38	1.80	1.47	6.75	3.75
t 1/2, h	1.56	—	0.98	8.90	—
MRT0-t, h	0.95	1.07	0.92	6.34	5.93
AUC0-inf, μg · h/mL or g	0.70	—	2.06	8.42	—

 The plain GFX solution data were obtained from the NDA application for the reported tissues: cornea, conjunctiva and aqueous humor. Parameters were estimated using a noncompartmental analysis from WinNonlin (ver. 1.5, Pharsight Inc.). Abbreviations are as described in Table 3, with the addition of MRT, mean residence time.

Multiple-Dose In Vivo Pharmacokinetic Study

The mean GFX content in various ocular tissues after instillation of one 50-μL drop of DPT-GFX (1.2% wt/vol) or Zymar (0.3% wt/vol) three times a day for 7 days (21 doses) is shown in Figure 9. In both the groups, corneal and conjunctival drug levels were higher than those in the aqueous humor. Compared with the cornea and conjunctiva, lower drug levels were observed in the sclera. In both the groups, drug levels were detectable until 24 hours in the cornea, conjunctiva, and aqueous humor, whereas scleral drug levels were not detectable beyond 12 hours. However, vitreous drug levels were detected until 12 hours in the DPT-GFX group, whereas no drug was detected after 6 hours in the Zymar group (Fig. 9b). GFX administration as DPT-GFX resulted in higher drug levels in the conjunctiva at all observed time points, when compared with that in the Zymar group (Fig. 9d). Similar results were obtained in the vitreous humor, except at the 6-hour time point. The mean pharmacokinetic parameters estimated for DPT-GFX and Zymar in various ocular tissues are summarized in Table 5. Higher drug levels (C _max) were achieved in all the tissues (∼1.3-fold in aqueous humor, cornea, and sclera; ∼2-fold in vitreous humor; ∼3-fold in conjunctiva) after DPT-GFX multiple doses when compared with Zymar (Table 5). Similarly, a larger AUC_0-∞ was observed after DPT-GFX was applied to the cornea (∼1.5-fold) and conjunctiva (∼3.5-fold), whereas similar levels were observed in the aqueous humor. Also, an increased half-life of Zymar was observed in the conjunctiva (∼2-fold) after DPT-GFX when compared with GFX in the contralateral eye. 

Figure 9.

 

GFX concentration in the (a) aqueous humor, (b) vitreous humor, (c) cornea, (d) conjunctiva, and (e) sclera of rabbits after multiple doses of one 50-μL drop of Zymar or DPT-GFX three times a day for 7 days (total, 21 drops). Each data point represents the mean ± SD (n = 3).

Table 5.

 

Pharmacokinetic Parameters Estimated for the Multiple-Dose Study after Topical Administration of One 50-μL Drop of DPT-GFX (1.2%) or Zymar (0.3%) Three Times a Day for 7 Days (Total, 21 Doses) to New Zealand White Rabbits

PK Parameter	Aqueous Humor	Vitreous Humor	Cornea	Conjunctiva	Sclera
DPT-GFX
C max, μg/mL or g	0.850 ± 0.146	0.370 ± 0.068*	21.630 ± 6.966	14.774 ± 8.172	0.0028 ± 0.0001
T max, h	0.417 ± 0.144	0.833 ± 0.289	0.417 ± 0.144	0.667 ± 0.289	0.25 ± 0.00
AUC0-t, μg · h/mL or g	5.429 ± 0.509	2.807 ± 0.957*	58.818 ± 11.362*	90.042 ± 32.798*	0.0126 ± 0.0016
t 1/2, h	13.082 ± 3.622	138.805 ± 222.84	6.688 ± 2.799	14.198 ± 1.778*	3.8679 ± 0.4271
AUC0-∞, μg · h/mL or g	8.987 ± 3.065	79.584 ± 129.82	63.352 ± 11.571*	136.517 ± 82.150	0.0139 ± 0.0022
Cl, 1 or g /h	0.215 ± 0.002†	0.082 ± 0.019	0.014 ± 0.003*	0.012 ± 0.006	14.7020 ± 1.5660
Fluctuation, %	182.845 ± 53.552	106.626 ± 38.723	501.367 ± 259.342	152.528 ± 100.221	136.977 ± 38.527
Accumulation index	2.897 ± 0.641	25.562 ± 40.161	1.782 ± 0.475	3.093 ± 0.316*	1.3138 ± 0.066
Zymar
C max, μg/mL or g	0.658 ± 0.029	0.178 ± 0.041	16.131 ± 7.017	5.087 ± 1.113	0.002 ± 0.001
T max h	0.333 ± 0.144	4.333 ± 2.887	0.417 ± 0.144	0.333 ± 0.144	0.667 ± 0.289
AUC0-t, μg · h/mL or g	4.846 ± 0.288	0.625 ± 0.477	36.785 ± 4.861	30.146 ± 1.177	0.009 ± 0.002
t 1/2, h	15.345 ± 3.225		7.132 ± 3.135	8.258 ± 1.729	6.607 ± 4.928
AUC0-∞, μg · h/mL or g	8.185 ± 0.454	40.630 ± 4.495	38.674 ± 5.833	0.019 ± 0.013
Cl, L or g /h	0.062 ± 0.008		0.005 ± 0.001	0.008 ± 0.001	16.563 ± 3.387
Fluctuation, %	151.702 ± 19.326		809.742 ± 789.072	179.850 ± 84.397	126.618 ± 44.762
Accumulation index	3.298 ± 0.576		1.857 ± 0.541	2.047 ± 0.299	1.784 ± 0.839
DPT-GFX/Zymar
C max, μg/mL or g	1.29	2.08	1.34	2.90	1.4
T max, h	1.25	0.19	1	2	0.37
AUC0-t, μg · h/mL or g	1.12	4.49	1.60	2.99	1.4
t 1/2, h	0.85		0.94	1.72	0.56
AUC0-∞, μg · h/mL or g	1.10		1.56	3.53	0.73
Cl, L or g /h	3.47		2.8	1.5	0.89

 Data are expressed as pharmacokinetic parameter estimates (mean ± SD; n = 3). Parameters were estimated using a noncompartmental analysis from WinNonlin (version 1.5, Pharsight Inc.). Abbreviations are as described in Table 3.

*   P < 0.05, significantly different from Zymar.

†   P < 0.001, significantly different from Zymar.

Discussion

The key findings of the present study are that (1) a g6 DPT formed a nanosized complex with GFX and increased its solubility by fourfold in the formulation studied; (2) the g6 DPT was capable of interacting with GFX through ionic, hydrogen bond, and hydrophobic interactions; (3) g6 DPT gained rapid entry into HCECs; (4) g6 DPT enhanced the permeability of GFX across bovine SCRPE layer; and (5) g6 DPT achieved higher drug concentrations in the target tissues after in vivo delivery in albino rabbits and maintained drug levels until 24 hours, potentially allowing a once-daily dose regimen. 

Drug solubility is a key limiting factor in the development of topical ophthalmic formulations and g6 DPT is capable of increasing GFX solubility by up to fourfold in the formulation tested in vivo. With an increase in drug solubility, there is a greater concentration gradient and hence a greater rate of drug transport into the intraocular tissues during the short precorneal residence of drops. Therefore, ophthalmic drugs are preferably prescribed in solution form as eye drops, to optimize target tissue concentrations. However, the intrinsic solubility of a compound, which depends on its physicochemical properties, dictates the maximum strength of an ophthalmic topical formulation. Many commercial eye drops provide drugs in solution at the high end of their aqueous solubility spectrum. If an increase in drug solubility is achieved, greater target tissue concentrations are likely, potentially resulting in enhanced and/or prolonged efficacy. Of note, the g6 DPT can enhance the solubility of GFX, a commonly used ophthalmic antibiotic. GFX, a zwitterionic compound with both acidic (pKa 6.03) and basic (pKa 8.38) groups, exhibited a pH-dependent solubility profile, with increasing solubility at lower pH in the absence of the dendrimer (Fig. 1a). At lower pH (5.65), the basic (NH) group remained predominantly ionized, contributing to increased solubility. Since the magnitude increase in solubility was higher in the pH 6.3 buffer (threefold), this buffer was selected for further increasing drug solubility with g6 dendrimer. With an increase in g6 dendrimer concentration, GFX solubility increased in a dose-proportionate manner (Fig. 1b). The solubility isotherm of GFX in g6 dendrimer indicated a linear relationship (A_L type curve) in the solubility. The slope of the observed isotherm was >1 (2.6981), indicating the formation of higher order complexes with more than one drug molecule interacting with one g6 dendrimer molecule.²³ With an increase in dendrimer concentration and the associated hydrochloric acid, the pH of the buffer decreased. When used at 15 mg/mL, the g6 dendrimer induced in a fourfold increase in solubility (12 mg/mL with dendrimer vs. 3 mg/mL with plain drug at pH 6) of GFX with a final formulation pH of 5.9, which is close to the pH of clinically available GFX eye drops (0.3% Zymar, pH 6.0). Thus, g6 DPT allowed the formulation of GFX as a solution at four times the strength of Zymar. Further, when stored at room temperature for more than 4 months, DPT-GFX was shelf-stable as a clear solution with no significant change in drug quantity, particle size, or pH. 

Association of GFX with g6 DPT is a likely explanation of the observed greater solubility of drug than that anticipated on the basis of a mere pH adjustment. To understand the nature of interactions and the binding forces involved in the formation of the drug–g6 dendrimer complex, ITC, FTIR, and NMR studies were performed. ITC is a powerful tool for investigating the thermodynamics of binding. ²⁴ In a typical ITC experiment, ligand (drug) is titrated against a molecule of interest, and the heat evolved or absorbed because of binding is converted into isothermic signals from which various thermodynamic parameters can be extracted. From a single ITC study, important thermodynamic parameters, including binding constant (Kb), enthalpy (ΔH), and entropy (ΔS) can be obtained. Based on ITC experiments, it can be inferred that electrostatic hydrogen bonding and hydrophobic interactions are involved in drug–dendrimer complex formation. The negative heat deflection of the binding reaction indicates an exothermic process. The exothermic binding process (Fig. 2a) is indicative of the involvement of electrostatic interactions in the binding process, ²⁵ possibly due to the interaction of the negatively charged carboxyl group of GFX with positively charged guanidine groups of g6 dendrimer. Further, titration in the presence of excess NaCl, a known competitive inhibitor of electrostatic (ionic) interactions, resulted in entropically favored binding in all three sites, indicating predominance of hydrophobic interactions on the addition of excess NaCl (Table 1). This result also confirms the presence of electrostatic interactions in the formation of the DPT-GFX complex, which were disrupted by excess NaCl. Further evidence of the ionic interaction between the carboxyl groups of GFX and the guanidine groups of the g6 dendrimer was confirmed from the FTIR studies (Fig. 3a). A negative enthalpy indicates enthalpically favored binding, and positive entropy indicates entropically favored binding. ²⁶ Enthalpically favored binding results from hydrogen bonding and vander Waal's interactions. Since the binding of GFX to the g6 dendrimer was both enthalpically and entropically favored at one site and entropically favored at two other sites (Table 1), hydrogen bonding and hydrophobic forces may be involved in the formation of the complex. Consistent with this, the ¹H-NMR study indicated the formation of hydrogen bonds, with the NH atoms in the guanidine end groups and amide linkages of the branching units in the g6 dendrimer serving as hydrogen bond donors and the electronegative O, N, and F atoms in the GFX serving as hydrogen bond acceptors (Fig. 3b). In the presence of 1 mM Triton X-100, enthalpically favored binding was observed at two sites (ΔH1 and ΔH3), indicating predominantly hydrogen bonding–based interactions (Fig. 2c). Triton X-100, a nonionic hydrophobic surfactant is expected to bind to the hydrophobic sites in the g6 dendrimer, thereby leaving GFX to bind predominantly through hydrogen bonding and ionic interactions. Also, the observed exothermic peaks in the presence of Triton X-100 indicated the presence of electrostatic interactions. All these results support the multiple binding sites for drugs in the g6 dendrimer and the involvement of ionic, hydrogen bonding, and hydrophobic interactions in the formation of drug–dendrimer complex (Fig. 4c). 

The in vitro uptake and permeability studies indicated the ability of the g6 DPT to rapidly enter the corneal epithelial cells (Fig. 6) and permeate the SCRPE (Fig. 5b). Apart from entrapping the drug and increasing its solubility, the g6 DPT also slowed the drug release (Fig. 5a), indicating its potential for sustained release. Fluoroquinolones, including gatifloxacin, are known to bind to melanin in ocular tissues. ²⁷ Greater drug release across a porous microdialysis membrane (∼70% or higher in 6 hours; Fig. 5a), when compared to permeability across the SCRPE (<0.8% at the end of 6 hours; Fig. 5b) suggests the ability of SCRPE to restrict drug transport possibly due to (1) tight junctions in RPE, (2) pigment binding of the drug in the choroid-RPE layer, and (3) thickness of SCRPE. Compared with plain drug, despite the lower free drug concentration on the donor side, the DPT-GFX allowed greater permeability across pigmented bovine SCRPE (Fig. 5b). This finding suggests that the dendriplex is much more permeable to the tissue than is the free drug. Permeability- and solubility-enhancing properties of the g6 DPT make it a suitable delivery system for overcoming barriers limiting ocular drug delivery. In addition to the advantages in enhancing the delivery of GFX to target tissues, DPT-GFX exhibited a two to four times faster killing rate in comparison to plain GFX solution against the resistant strain of S. aureus (MRSA), which is commonly found in the conjunctiva (Table 2). Thus, both the in vitro delivery and efficacy studies strongly support the advantage of using g6 DPT as an ocular delivery vehicle. 

The superiority of g6 DPT in improving ocular drug delivery was further supported by the in vivo study results. After single-dose administration of DPT-GFX, higher AUC_0-∞ was observed in the cornea and conjunctiva in comparison to 0.3% GFX solution. The AUC_0-∞ increased by ∼2-fold in the cornea and by ∼13-fold in the conjunctiva. Also, the half-life of GFX increased from 0.65 hour to 3.44 hours in the conjunctiva after DPT-GFX administration. In addition, DPT-GFX resulted in a 9- and 32-hour half-life in the aqueous humor and vitreous humor, respectively. Since the increase in GFX dose from 0.3% to 0.5% did not cause any proportionate increase in any of the pharmacokinetic parameters (Table 4), the higher bioavailability of GFX after a dose of DPT-GFX may not be primarily due to the higher drug content (1.2%). The cell-penetrating property of g6 DPT, as evident in the HCEC uptake study (Fig. 6), may contribute to the increased bioavailability of GFX after DPT-GFX administration. Since other tissues were not collected in the NDA study of GFX, comparison cannot be made with DPT-GFX levels in other ocular tissues. Although the DPT-GFX in the single-dose study was instilled in a single eye and compared with NDA data for GFX, the multiple-dose study involved application of DPT-GFX in one eye and Zymar in the other. Similar to the single-dose study, DPT-GFX was superior in the multiple-dose study. However, the change ratios were lower in the multiple-dose study for the cornea and conjunctiva. Further, the half-lives were prolonged following multiple doses of DPT-GFX as well as GFX-treated eyes for all tissues in which half-life could be estimated. A possible reason for these observations in the multiple-dose study is the delivery of the drug to the contralateral eye via the systemic circulation. ²⁸ Specifically, from the eye dosed with 1.2% wt/vol strength of DPT-GFX, it can be anticipated that there is a significant drug contribution to the contralateral eye treated with a lower dose of 0.3% wt/vol Zymar. The higher bioavailability and increased half-life with DPT-GFX in single-dose as well as multiple-dose studies clearly demonstrates the superior drug delivery ability of DPT. 

Successful eradication of bacterial ocular infections requires delivery of efficacious concentrations at the target sites for prolonged periods. The efficacy of antibiotics may be predicted based on their pharmacokinetic–pharmacodynamic parameters. ²⁹ For fluoroquinolones, a C _max/MIC₉₀ ratio >10 is necessary for clinical and microbiologic success and to limit the development of bacterial resistance. Similarly, an AUC/MIC₉₀ >30 is needed for maximum bacterial killing. ³⁰ The C _max/MIC₉₀ and AUC/MIC₉₀ ratios were greater than 10 and 35, respectively, for DPT-GFX after acute single doses (2 drops) and multiple doses (21 drops over 7 days) in both cornea and conjunctiva, which are the target tissues for corneal ulcer and conjunctivitis (Table 6). The MIC level for GFX to eradicate bacterial endophthalmitis isolates ranges from 0.09 to 0.38 μg/mL. ⁸ DPT-GFX maintained drug levels in the target range for 24 hours after a single bilateral dose (Fig. 8c). Similarly, after multiple doses, DPT-GFX achieved the target concentration in the vitreous humor for 12 hours, whereas no drug levels were detected in the vitreous at 12 hours after administration of Zymar (Fig. 9b). All these in vivo results clearly indicate the advantage of using DPT-GFX for enhancing drug delivery to the eye. 

Table 6.

 

Pharmacokinetic-Pharmacodynamic Determinants of Efficacy for GFX after Acute and Multiple-Dose Topical Administration of DPT-GFX in New Zealand White Rabbits

Ocular Isolates	MIC90	C max/MIC				AUC/MIC
		Cornea		Conjunctiva		Cornea		Conjunctiva
		Acute Dose	Multiple Dose	Acute Dose	Multiple Dose	Acute Dose	Multiple Dose	Acute Dose	Multiple Dose
S. aureus (S), n = 25	0.22	120	98	47	67	283	288	161	621
S. pneumoniae, n = 20	0.25	106	87	41	59	249	253	142	546
S. viridans, n = 20	0.38	70	57	27	39	164	167	93	359
P. aeruginosa (S), n = 25	0.38	70	57	27	39	164	167	93	359
S. marscens, n = 10	0.38	70	57	27	39	164	167	93	359
Hemophilus sp., n = 10	0.064	414	338	162	231	972	990	555	2133
Moraxella sp., n = 10	0.032	828	676	323	462	1945	1980	1109	4266
All susceptible, n = 170	0.24 ± 0.15	110	110	43	62	259	264	148	569

  C _max and AUC were obtained from the current study and the MIC₉₀ for various ocular pathogens was obtained from References 8, 29, and 30.

In summary, g6 DPT, a novel dendrimer, enhances aqueous solubility of GFX, thereby providing a drug concentration (1.2% wt/vol) much higher than what is currently feasible. The g6 DPT forms isotonically stable complexes of nanosize with GFX through ionic, hydrogen bond, and hydrophobic interactions. The DPT-GFX formulation is shelf-stable and has an acceptable pH of 5.9 suitable for ocular administration. DPT-GFX enhances delivery of GFX to cornea and conjunctiva with favorable C _max/MIC and AUC/MIC after a single dose, potentially allowing decreased frequency of administration. In addition to the advantages in enhancing the delivery of GFX to target tissues, DPT-GFX exhibited a 2- to 4-times faster killing rate in comparison to GFX solution against the resistant strain of S. aureus (MRSA), which is commonly found in the conjuctiva. Thus, both the in vitro delivery and efficacy studies strongly support the advantage of using g6 DPT as an ocular delivery vehicle. 

Footnotes

 Supported by Visionary Therapeutics Corporation, Richmond, VA.

Footnotes

 Disclosure: C. Durairaj, Visionary Therapeutics Corporation (R); R.S. Kadam, None; J.W. Chandler, Visionary Therapeutics Corporation (C); S.L. Hutcherson, Visionary Therapeutics Corporation (E); U.B. Kompella, Visionary Therapeutics Corporation (F, C)

The authors thank Chris Pillar (Eurofins Medinet) for the microbiology work and Ashish Thakur and Namdev Shelke for their assistance during the animal studies.