Design and Optimization of Ciprofloxacin Hydrochloride Biodegradable 3D Printed Ocular Inserts: Full Factorial Design and In-vitro and Ex-vivo Evaluations: Part II

Abstract

Coupling hot-melt extrusion (HME) with fused deposition modeling three-dimensional printing (FDM-3DP) can facilitate the fabrication of tailored, patient-centered, and complex-shaped ocular dosage forms. We fabricated ciprofloxacin HCl ocular inserts by coupling high-throughput, solvent-free, and continuous HME with FDM-3DP. Insert fabrication utilized biocompatible, biodegradable, bioadhesive Klucel^™ hydroxypropyl cellulose polymer, subjected to distinct FDM-3DP processing parameters, utilizing a design of experiment approach to achieve a tailored release profile. We determined the drug content, thermal properties, drug-excipient compatibility, surface morphology, in vitro release, antibacterial activity, ex-vivo transcorneal permeation, and stability of inserts. An inverse relationship was noted between insert thickness, infill density, and drug release rate. The optimized design demonstrated an amorphous solid dispersion with an extended-release profile over 24 h, no physical or chemical incompatibility, excellent mucoadhesive strength, smooth surface, lack of bacterial growth (Pseudomonas aeruginosa) in all release samples, and prolonged transcorneal drug flux compared with commercial eye drops and immediate-release inserts. The designed inserts were stable at room temperature considering drug content, thermal behavior, and release profile over three months. Overall, the fabricated insert could reduce administration frequency to once-daily dosing, affording a promising topical delivery platform with prolonged antibacterial activity and superior therapeutic outcomes for managing ocular bacterial infections.

Graphical Abstract:

1. Introduction

Three-dimensional printing (3DP), also known as additive manufacturing, has been widely regarded as an industrial technology of the twenty-first century. Over the past 20 years, this technology has been employed in a range of new fields, including healthcare. It is used to create customized prostheses, dental implants, and, more recently, pharmaceutical applications (Agarwal et al., 2022). According to recent studies published in the literature, rapid prototyping of personalized products, designed using computer-aided design (CAD) software or imaging methods, can be achieved using fused deposition modeling (FDM) three-dimensional printing (3DP). FDM-3DP is expected to transform the pharmaceutical sector owing to its capacity to produce small quantities of customized pharmaceuticals with unique designs, drug combinations, and customized release characteristics (Awad et al., 2018). Extending drug residence time or elevating corneal penetration are crucial objectives for developing ocular formulations. To afford improved ocular bioavailability, the design and development of novel formulations could overcome the drawbacks of the current ocular topical products.

Bacterial keratitis (BK) is a serious corneal inflammation caused by bacteria, which can lead to severe visual disability without proper treatment (Youssef et al., 2021). The prevalence rate of BK in the United States (US) is estimated at 200,0000 cases per year (Thakkar et al., 2021). BK remains one of the most common causes of irreversible blindness worldwide (Al-Mujaini et al., 2009). The causative pathogens of BK vary mainly based on geographical location and climate (Thakkar et al., 2021). Gram-positive Staphylococcus aureus and Staphylococcus epidermidis are predominant pathogens isolated in the northern US; however, gram-negative Pseudomonas aeruginosa is the most common causative agent in the southern US (Sand et al., 2015; Thakkar et al., 2021). BK can also be caused by pathogens such as gram-positive Streptococcus and Bacillus species and gram-negative bacteria, such as Moraxella lacunata, Serratia marcescens, Haemophilus influenzae, and Microbacterium liquefaciens (Bertino, 2009; Pachigolla et al., 2007; Teweldemedhin et al., 2017). Effective early management of BK can prevent progression to corneal perforation, ulceration, endophthalmitis, or blindness (Bertino, 2009; Pachigolla et al., 2007).

In the US, monotherapy with fluoroquinolones, such as ciprofloxacin (CIP), moxifloxacin, gatifloxacin, and levofloxacin, has been established as the first choice for BK treatment (Bertino, 2009), owing to its broad-spectrum activity, high tissue penetration, and ocular safety and toxicity profiles (Wozniak and Aquavella, 2017). CIP is a second-generation, broad-spectrum, bactericidal antibiotic, exhibiting the highest activity against gram-negative P. aeruginosa considering available quinolones (Thai et al., 2022). CIP is well-known to target topoisomerases II and IV, which are involved in bacterial DNA synthesis and replication. CIP is available as an eye drop and ointment (0.3% w/v CIP base) for ocular applications. However, commercial eye drops warrant frequent application, given their low ocular bioavailability. Only 1–5% of the drug applied to the eye surface penetrates the intraocular tissues, given the precorneal elimination of the remaining amount through nasolacrimal drainage, high tear fluid turnover, and overflow from the conjunctival sac (Farkouh et al., 2016; Youssef et al., 2021). Although ointments can prolong the contact time between CIP and the ocular surface and thus improve ocular bioavailability, these formulations are frequently associated with numerous side effects, including blurred vision, tearing, itching, redness, greasiness, eye discomfort, and dryness (Patel et al., 2013;). Therefore, both ophthalmic products could reduce patient compliance and convenience, thereby resulting in ineffective therapy.

Following placement in the conjunctival sac, ocular inserts can prolong the contact time between the therapeutic agent and ocular surface (Khan et al., 2022; Thakkar et al., 2021). These inserts could reduce application frequency, precorneal elimination, and systemic exposure, along with improved shelf life and bioavailability when compared with solution eye drops (Thakkar et al., 2021). In addition, ocular inserts can be designed with various characteristics by selecting the appropriate polymer composition to achieve desired features such as biocompatibility, mucoadhesivity, and biodegradability to overcome the potential drawbacks associated with conventional inserts (Thakkar et al., 2021). For example, the most commonly available ocular insert, Ocusert^®, is fabricated using biocompatible synthetic polymers; however, the insert is non-biodegradable, necessitating removal after application (Khan et al., 2022).

Polymeric matrices for ocular inserts can be prepared using two methods: solvent casting and melt casting (Khan et al., 2022). According to a literature review, solvent casting is primarily used to prepare the majority of ocular inserts (Thakkar et al., 2021). For solvent casting, the drug and polymer are initially dissolved in a suitable solvent, followed by casting into the desired shape and drying to remove the organic solvent (Khan et al., 2022). However, this method has several shortcomings, including prolonged time required for solvent removal, limited large-scale production, and safety issues owing to traces of organic solvents (Khan et al., 2022; Thakkar et al., 2021). Thermal processing starts by mixing active and inactive ingredients, followed by thermo-mechanical shaping by extrusion, injection, or compression molding (Khan et al., 2022). Thermal methods afford several advantages over solvent casting methods, such as solvent-free, continuous, large-scale production, and low-cost techniques (Khan et al., 2022; Thakkar et al., 2021).

Hot-melt extrusion (HME) is a successful, versatile, and continuous thermal process, which has been employed in the plastics, rubber, food, and pharmaceutical industries (Patil et al., 2016). HME is mainly used for preparing amorphous solid dispersions to improve the solubility of poorly soluble drugs and, thus, improve bioavailability via different routes of administration, including oral (Narala et al., 2022), ocular (Thakkar et al., 2021), and dermal (Shettar et al., 2021). To the best of our knowledge, two HME-approved ocular inserts are currently available, Lacrisert^® topical insert for relief from moderate to severe dry eye symptoms and Ozurdex^® intravitreal dexamethasone implant for macular edema, diabetic macular edema, and noninfectious uveitis, with several under development (Simões et al., 2019).

3DP has been used to create customized pharmaceutical dosage forms for diverse applications (Agarwal et al., 2022). Coupling HME with FDM-3DP can allow the production of customized patient-centered pharmaceuticals with unique designs, complex shapes, and tailored release characteristics for ocular applications.

We have previously developed extended-release CIP-HCL-HME ocular films to enhance ocular surface retention, improve ocular bioavailability, and reduce administration frequency to once-daily dosing, thereby improving patient compliance and providing better treatment outcomes (Alzahrani et al., 2022a). Subsequently, the lead HME film was transformed into a filament and printed using an FDM-3D printer; however, the formulation failed to preserve the primary objective of the study, releasing more than the target release value (60%) during the first 6 h of the release study (83%) (Alzahrani et al., 2022a). The current study aimed to extend our previous studies and explore the feasibility of coupling HME with FDM-3DP technology for developing customized sustained-release CIP-HCL inserts for ocular applications. In addition, we examined the effect of different printing process parameters on distinct critical quality attributes of formulated inserts, including drug release, using a design of experiment (DOE) approach. This additive manufacturing technique could overcome the shortcomings of conventional ophthalmic dosage forms, as well as the drawbacks of earlier rudimentary approaches, such as solvent casting methods for ocular insert fabrication.

2. Materials and methods:

2.1. Materials

CIP-HCl (Purity: ≥95%) was acquired from Cayman Chemical Company (Cayman Chemicals, Ann Arbor, MI, USA). Klucel^™ hydroxypropyl cellulose (HPC) JXF grade was a generous gift from Ashland^™(Ashland Inc., Bridgewater, NJ, USA). POLYOX^™ WSR N10 (PEO N10) was a kind gift from Colorcon, Inc. (Irvine, CA, USA). Polyethylene glycol 4000 (PEG 4000) was purchased from Fischer Chemicals (Hampton, NH, USA). Parteck^® SI 150 (sorbitol) was a kind gift from Millipore Sigma (Burlington, MA, USA). Other chemicals, such as solvents and reagents, were of analytical grade and were procured from Fisher Scientific (Fair Lawn, NJ, USA). The centrifuge tubes (50 mL), high-pressure liquid chromatography (HPLC) vials, and glass vials were purchased from Fischer Scientific (Hampton, NH, USA). Millex^® syringe nylon membrane filters (0.45 μm) were purchased from MilliporeSigma (St. Louis, MO, USA).

2.2. Biological tissues and samples

The whole eye globe male and female albino New Zealand rabbits (weight, 4.75-5.75 lbs; age, 8-12 weeks) were acquired from Pel-Freez Biologicals (Rogers, AR, USA). The eyes were delivered overnight in Hanks' balanced salt solution and maintained under cooled conditions throughout the shipping process. On arrival at the laboratory, corneas were excised and permeation experiments were performed. Microbial strains were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA).

2.3. Methods

2.3.1. HPLC

According to the CIP-HCL USP monograph, CIP-HCL was quantified using a validated reversed-phase HPLC method. The experiment was performed using a Waters HPLC system (Waters Corp., Milford, MA, USA) fitted with an autosampler, UV/VIS detector, and Empower software. Waters Symmetry^® C18 column (150 mm × 4.6 mm × 5.0 μm), with a detection wavelength (λmax) of 278 nm, was used for chromatographic separation. Acetonitrile and phosphoric acid solution (25 mM; pH of 3.0 ± 0.1) were combined in a 13:87 (% v/v) ratio and pumped isocratically through the detector. The pH was regulated using triethylamine. The flow rate was set at 1.2 mL/min, and the run time was set to 10 min. The method was linear over the concentration range of 1-100 μg/mL of CIP-HCL.

2.3.2. Preparation of the physical mixture (PM)

Table 1 presents the composition of the immediate and optimized (OPT) sustained-release CIP-HCL ocular inserts prepared by HME in our earlier investigation. The polymers (PEO N10 and/HPC), plasticizers (sorbitol and PEG 4000), and drugs were sieved through a U.S. mesh (#30). Subsequently, ingredients were mixed using a geometric dilution technique. Next, all blends were mixed using a V-shaped blender (MaxiBlend^™, GlobePharma, NJ, USA) at 25 rpm for 10 min to ensure homogenous PM. The powder was then sealed in 5 mL polyethylene bags and placed in a desiccator at room temperature for future studies.

Table 1.

Composition of the immediate and sustained release ciprofloxacin hydrochloride 3DP ocular inserts.

Formulation	Ingredients (% w/w)
	CIP-HCL	HPC JXF	PEO N10	PEG 4000	Sorbitol	Batch size (g)
CIP-3DP-SR	6	91	0	3	0	10
CIP-3DP-IR	6	22.375	69.375	0.75	1.5	10

2.3.3. Experimental Design

Based on accumulated drug delivery literature, controlling the size and shape of the 3DP dosage forms by controlling 3DP printing parameters can affect the drug dissolution rate (Obeid et al., 2021). Therefore, we attempted to build a design space to examine the relationship between specific 3DP processing parameters and CIP release from FDM-3DP ocular inserts, using a 2³ full factorial experimental design. Among several factorial models, we selected a randomized regular two-level design to study the interaction between 3DP parameters (independent variables) and drug release (dependent variable). Three independent factors (variables), including the insert thickness (X₁, mm), insert infill (X₂, %), and layer height (X₃, mm), were selected and studied at two different levels: level −1 (X₁: 0.3 mm, X₂: 75 %, and X₃: 0.1 mm) and level +1 (X₁: 1.0 mm, X₂: 100 %, and X₃: 0.4 mm), as shown in Table 2. The levels of independent variables were chosen based on a literature review (X₂ and X₃) and FDA-marketed ocular insert dimensions (X₁) (Saettone and Salminen, 1995; Vo et al., 2020; Wang et al., 2020; Zhang et al., 2020). The ideal thickness of ocular inserts should range between 0.3 and 1.0 mm for improved patient compliance (Saettone and Salminen, 1995). All three processing parameters were randomly displayed in the Design-Expert^® software (StatEase Inc., version 13, Minnesota, USA) and represented in eight experimental runs, as shown in Table 3. The response variable selected for optimizing the 3DP ocular inserts was the % drug release at 0.5 h (Y₁) and 6 h (Y₂).

Table 2.

Independent and dependent variables of the 2-level randomized factorial design.

Independent factors	Levels
	−1 (Low)	+1 (High)
X1 – insert thickness (mm)	0.3	1.0
X2 – insert infill (%)	75	100
X3 – layer height (mm)	0.1	0.4
Dependent factors
Y1 – release at 0.5 h (%)	Y2 – release at 6 h (%)

Table 3.

Experimental design layout as per the 2-level randomized factorial design.

Formulation	Run	Assigned independent variables			Actual independent variables			Response
		X1	X2	X3	Insert thickness (mm)	Insert infill (%)	Layer height (mm)	Release at 0.5 h (%)	Release at 6.0 h (%)
CIP-3DP-SR	1	+1	+1	−1	1	100	0.1	35.9	70.8
	2	−1	+1	+1	0.3	100	0.4	49.3	79.8
	3	+1	+1	+1	1	100	0.4	35.1	72.1
	4	−1	+1	−1	0.3	100	0.1	52.3	81.3
	5	+1	−1	+1	1	75	0.4	45.2	87.8
	6	−1	−1	−1	0.3	75	0.1	62.8	100
	7	+1	−1	−1	1	75	0.1	42.5	85.3
	8	−1	−1	+1	0.3	75	0.4	59.9	100

3DP, three-dimensional printing; CIP, ciprofloxacin; SR, sustained release.

2.3.4. HME process

HME was performed using a Haake Minilab (Thermo Fisher Scientific) co-rotating twin-screw extruder with a screw speed of 50 rpm. PMs (8.0 g) were fed into the preheated barrel (140°C) at a steady rate through the hopper. A cylindrical die (2.0 mm) was used to extrude PMs into filaments (diameter, ~ 2 mm). To prepare extruded filaments exhibiting a suitable diameter (~1.7 mm) for fitting to FDM-3DP, the filaments were manually tuned by tugging them while leaving the extrusion die using a Thermo Fisher^™ Scientific digital caliper (0-150 mm) (Fig. 1). The extruded filaments were cooled, sealed in plastic bags with silica packs, and stored at room temperature (25 °C) for future studies.

Figure 1.

Manual tugging for the HME extruded filament to adjust diameter for FDM-3DP.

2.3.5. Three-Point Bend Test (Repka-Zhang test).

Before FDM-3DP, the Repka-Zhang test was performed to evaluate the mechanical properties (flexibility and brittleness) of HME extruded filaments (Zhang et al., 2017). The brittleness and flexibility of extruded filaments were evaluated using a TA-XT2i texture analyzer (Stable Micro Systems, Texture Technologies Corporation, London, UK), equipped with a TA-95N 3-point bend probe set. The test was performed using the following parameters: supporting gap, 25 mm; pre-test speed, 2 mm/s; test speed, 10 mm/s; post-test speed, 10 mm/s; probe moving distance, 10 mm; trigger force, 10 g. Briefly, extruded filaments were collected and chopped into small segments (50 mm in length). Each segment was placed on a sample holder perpendicular to the knife blade above the supporting ridges (Fig. 2). The knife blade was lowered at a speed of 10 mm/s until it reached 10 mm below the segment; polylactic acid (PLA) was used as reference material. Data were collected and analyzed using the Exponent software (version 6.1.5.0, Stable Micro Systems, Godalming, UK). Five segments of each filament were examined.

Figure 2.

3-point bend test using TA-XT2i texture analyzer.

2.3.6. FDM-3DP

The oval 3D printed inserts were designed considering different thicknesses (0.3 or 1.0 mm), layer height (0.1 or 0.4 mm), and infills (75 or 100 %) using Autodesk Tinkercad, an online simulator modeling tool (Tinkercad, Autodesk Inc., CA, USA) and saved as STL format files for printing. A fused filament fabrication 3D printer (Prusa i3 3D desktop printer, Prusa Research, Prague, Czech Republic) was used to print the oval-shaped inserts, and Cura^© software (version 15.04; Ultimaker, Geldermalsen, Netherlands) was employed for the 3DP process to adjust printing parameters (Fig. 3). The printing temperature was set at 180°C, whereas the bed temperature was 40°C, and the inserts were printed at a speed of 10 mm/s.

Figure 3.

A schematic graph demonstrating the shape, thickness, and view of the ciprofloxacin hydrochloride three-dimensional printed ocular insert. Abbreviations: CIP, ciprofloxacin; FDM-3DP, fused deposition modeling three-dimensional printing.

2.3.7. Control formulation

CIP-3DP-IR was used as a control in ex vivo permeation and in vitro release experiments.

CIP hydrochloride ophthalmic solution control (CIP-C)

Ophthalmic solution eyedrops (Ciprofloxacin Ophthalmic Hydrochloride solution, 0.3% as a base, Leading Pharma LLC, NJ, USA; Lot # 018B041) were used as a control formulation for the ex vivo permeation and antibacterial testing studies.

2.3.8. Characterization of ocular inserts

2.3.8.1. Thickness and weight

The CIP-3DP-SR (eight runs) and CIP-3DP-IR inserts were analyzed using a VWR^® digital caliper (VWR International Company, Radnor, PA, USA). The weight of each insert was recorded using a calibrated Mettler Toledo excellence balance (Mettler Toledo, Columbus, OH, USA). All thickness and weight measurements were performed in triplicates.

2.3.8.2. Surface pH

The 3DP inserts were placed in glass vials (3.0 mL) for surface pH measurements. Isotonic phosphate-buffered saline (IPBS; pH 7.4) was added to the vials. The pH was measured using a Mettler Toledo pH meter (FiveEasy^™, Columbus, OH, USA) equipped with an Inlab^® Micro Pro-ISM probe.

2.3.8.3. Drug content

The molded ocular inserts (~5 mg) were dissolved in a mixture of acetonitrile and 25 mM phosphoric acid solution (pH 2.0 ± 0.1, adjusted with TEA), mixed in a ratio of 13:87 (% v/v) in a volumetric flask (10 mL). The extract was vortexed for 3 min at 2000 rpm and sonicated (Bransonic^® ultrasonic cleaner, USA) for 5 min. The mixtures were then centrifuged using AccuSpin^™ 3R (Fisherbrand^™, Waltham, MA, USA) at 13,000 rpm for 15 min. The supernatant was passed through a nylon membrane filter (0.45-μm pore size) and analyzed for CIP-HCL content using the HPLC method described in subsection 2.3.1.

2.3.8.4. Swelling index (SI)

To determine the SI of 3DP inserts, inserts were weighed (initial weight) individually and placed into glass vials (5 mL) containing IPBS (pH 7.4; 34 ± 0.2°C). The insert was removed at regular time intervals for up to 30 min, excess IPBS media was removed using filter paper, and the insert was reweighed. SI (%) was calculated using the following equation:

$$\text{Swelling index}(\%)=\frac{\text{Initial weight of the insert}(\text{mg})\text{−}\text{Weight of swollen insert (mg)}}{\text{Initial weight of the insert (mg)}}\times 100$$

2.3.8.5. Moisture absorption and loss (%)

The initial weights of the three CIP-3DP-SR-OPT inserts were recorded, and inserts were placed in a desiccator with aluminum chloride (absorption) or anhydrous calcium chloride (loss) for 3 days. After the testing period, the inserts were collected and reweighed, and the % moisture absorption and loss were calculated by applying the following equations:

$$\begin{gathered} \text{Moisture absorption}(\%)=\frac{(\text{Final weight}\text{−}\text{Initial weight})}{(\text{Initial weight})}x100 \\ \text{Moisture loss}(\%)=\frac{(\text{Final weight}\text{−}\text{Initial weight})}{(\text{Initial weight})}\mathrm{x}100 \end{gathered}$$

2.3.9. Differential scanning calorimetry (DSC)

Samples (~6 mg) of pure CIP-HCL, pure excipients, CIP-3DP-IR (PM, filament, and insert), and CIP-3DP-SR-OPT (PM, filament, and insert) formulations were hermetically sealed in aluminum pans and analyzed by DSC (Discovery 25; TA Instruments, New Castle, DE, USA) in a single heating ramp, with an empty pan as a control reference. Samples were scanned from 25 to 200°C at a heating rate of 10°C/min under an inert nitrogen atmosphere (50 mL/min).

2.3.10. Fourier transform infrared (FTIR) spectroscopy

The FTIR spectra of pure CIP-HCL, pure excipients, CIP-3DP-IR (PM, filament, and insert), and CIP-3DP-SR-OPT (PM, filament, and insert) were collected using a Cary-630 FTIR spectrometer (Agilent Technologies, Santa Clara, CA, USA) equipped with a MIRacle attenuated total reflection (Pike Technologies MIRacle ATR, Madison, WI, USA), fitted with a single-bounce, diamond-coated ZnSe internal reflection element (Butreddy et al., 2020). Samples were placed on the surface of the crystal and pressed to obtain a uniform solid-crystal contact using an instrument built-in pressure tower. The FTIR spectra were collected in transmittance mode between 650–4000 cm⁻¹, with 16 scans and 4 cm⁻¹ resolution.

2.3.11. Surface morphology studies– scanning electron microscopy (SEM)

The surface morphology of the pure drug (CIP-HCl) crystals, PM, HME filaments, and drug-loaded 3DP insert samples was examined using a JSM-7200FLV scanning electron microscope (JOEL, Peabody, MA, USA), with an accelerating voltage of 5 kV. All samples were placed on double-adhesive tape and attached to the SEM stubs. Subsequently, samples were sputter-coated with platinum in an argon environment using a fully automated Denton Desk V TSC Sputter Coater (Denton Vacuum, Moorestown, NJ, USA) before imaging.

2.3.12. In vitro release

Accurately weighed 3DP inserts were placed above the bottom of the glass vials (20 mL). A stainless-steel US mesh (#10) was placed above the inserts with a magnetic stirrer (13.0 mm [length] and 3.0 mm [diameter]) above the mesh screen. The vials were filled with PBS (pH 7.4, 20 mL) containing 2.5% randomly methylated-β-cyclodextrin (RMβCD) and maintained at a 34 ± 0.2°C under continuous magnetic stirring on a 10 IKA magnetic stirrer (IKA Works, Inc., Wilmington, NC, USA). Aliquots (1.0 mL) were removed at regular intervals and replaced with an equal volume of freshly prepared release medium. The release medium was selected based on our previous investigations (Youssef et al., 2021). CIP-3DP-IR inserts composed primarily of PEO were used as controls. The amount of CIP in all collected aliquots was analyzed using the HPLC method described in subsection 2.3.1. The release data were fitted to the five main release models using the DDSolver software for modeling and comparing drug dissolution profiles, as shown in Table 4.

Table 4.

Models and their corresponding equations for release data fitting.

Model	Equation
Zero-order	M0-M = k.t
First-order	ln M = k.t
Higuchi	M0-M = k.t1/2
Hixson–Crowell	M01/3-M1/3 = kt
Korsmeyer-Peppas	log (M0-M) = n log t + log k

Where M_o and M represent initial drug content at the time t_o and drug content remaining at time t, respectively; Zero-order model: % drug released vs. time; First order model: Amount drug remaining vs. time; Higuchi model: % drug released vs. square root of time; Korsmeyer-Peppas model: log % drug released vs. log time.

2.3.13. Evaluation of antibacterial activity

In the collected release samples, the antibacterial activity of CIP-HCL was tested against P. aeruginosa (ATCC BAA-2018) from the American Type Culture Collection (ATCC, Manassas, VA, USA). Antimicrobial susceptibility testing was performed in accordance with a modified version of the Clinical & Laboratory Standards Institute protocol (CLSI, 2012). Cation-adjusted Mueller-Hinton broth was used to dilute samples (pH 7.3). Diluted samples (10 μL) were then transferred to microdilution plates (96 wells). As per the CLSI procedure, inocula were prepared by adjusting the OD630 of bacterial suspensions in the incubation broth to provide appropriate inocula. The final assay plate was loaded with 5% Alamar Blue^™. The CIP-HCL solution in the release medium was used as the positive control, whereas IPBS (pH 7.4) containing 2.5% RMβCD was used as the negative control, given that it was used as the release medium. The optical density was measured for each panel well using a Bio-Tek plate reader at 544ex/590em before and after incubation at 35°C for 24 h. The minimum inhibitory concentration (MIC) was determined for all tested formulations and defined as the lowest test CIP-HCL concentration that resulted in no visual growth. All experiments were performed in triplicate.

2.3.14. Ex vivo permeation studies

The permeation study was performed using freshly excised rabbit corneas with PermeGear vertical glass diffusion cells with spherical joints to preserve the real curvature of corneas (PermeGear^® Inc., Hellertown, PA, USA). The corneas were fixed between the donor and receiver compartments, with the epithelial surface facing the donor compartment containing the test or control formulations with simulated tear fluid (STF; 10 μL). The STF (pH of 7.4 ± 0.2) was prepared by dissolving calcium chloride (0.0084%), sodium chloride (0.678%), potassium chloride (0.138%), and sodium bicarbonate (0.218%) in Milli-Q water (Stjernschantz and Astin, 1993). Both compartments were clamped together using stainless-steel PermeGear clamps. The receiver compartment was filled with freshly prepared IPBS (pH 7.4, 5.0 mL) containing 2.5% RMβCD. The vertical Franz diffusion cells were maintained under continuous stirring, and the temperature was maintained at 34 ± 0.2°C using a circulating heater. Commercial eye drops and CIP-3DP-IR inserts were used as controls. The effective corneal area (A) available for drug diffusion was 0.636 cm^2, and the experiment was conducted for 3 h. Samples (0.5 mL) were withdrawn from the receiver compartment at predetermined time points and replaced with an equal volume of freshly prepared receiver medium. The amount of drug in the receiver compartment was quantified using the HPLC method described in subsection 2.3.1. The rate of transcorneal permeation (dM/dt) was calculated from the slope of the plot of the amount permeated versus time. The steady-state flux (J_ss) was calculated from the rate of transcorneal permeation (dM/dt) and the effective permeation area using the following equation:

$${\mathrm{J}}_{\text{ss}}=\frac{\mathrm{d}\mathrm{M}∕\mathrm{d}\mathrm{t}}{\mathrm{A}}$$

2.3.15. Ex vivo mucoadhesive testing

Freshly excised rabbit corneas were used to evaluate the mucoadhesive force and work of adhesion (WOA) of the fabricated 3DP ocular inserts (Fig. 4). Mucoadhesion was examined using a TA.XT2i texture analyzer, equipped with a 50 N load cell and a TA-57R (diameter: 7 mm) probe. The ocular insert was adhered to double-face adhesive tape attached to the base of the testing probe, which was fixed to the mobile arm of the texture analyzer. A tissue holder was loaded in a glass beaker (500 mL) filled with IPBS (pH 7.4) and maintained at 34.0 ± 0.5°C. The corneas were then fixed between the two parts of the tissue holder, with the epithelial surface facing the probe. STF (10 μL) was added to the ocular surface using an Eppendorf^® pipette. The following parameters were set; applied force, 5.0 N; pre-test speed, 0.1 mm/s; test speed, 1.0 mm/s; pre-test speed, 0.1 mm/s; contact time, 60 and 180 s. The maximum force necessary to detach the probe from the ocular surface (Fmax) was determined and used to calculate the WOA using the Exponent software. All measurements were performed in triplicates.

Figure 4.

Mucoadhesive strength measurement of ciprofloxacin hydrochloride fused deposition modeling 3D printed insert using excised rabbit corneal with TA.XT2i. Abbreviation: IPBS, Isotonic phosphate-buffered saline.

2.3.16. Stability studies

The stability of the optimized design of the CIP-3DP-SR-OPT insert was evaluated at room temperature (25 ± 2°C, 60% RH ± 5% RH) for up to three months of storage (final time point assessed). Briefly, the inserts were weighed before placement in closed glass vials (3 mL) and stored at 25 ± 2°C. The inserts were evaluated for changes in drug content, thermal behavior, and release characteristics upon storage at regular time intervals.

2.3.17. Data analysis

Design-Expert software (version 13; Stat-Ease, Inc., Minneapolis, MN, USA) was used to optimize the 3DP process parameters. Statistical analysis was performed using the statistical function of Microsoft^® (MS) Office Excel (Office365 2019, Microsoft Corporation, Redmond, WA, USA).

3. Results and Discussion

3.1. Filament preparation by HME

Herein, PMs of the sustained release (CIP-3DP-SR) and immediate-release (CIP-3DP-IR) inserts (Table 1), developed and optimized in our earlier investigation, were extruded as filaments rather than films. A Haake Minilab co-rotating twin-screw extruder was used to prepare the FDM-3D printer ink (filaments) to conserve materials in small batches (~8.0 g). The polymer grades used in the present study (HPC JXF and PEO N10) provided excellent flowability and processability via HME; hence, the feeding rate was constant (Thakur and Thakur, 2015). Some statistically insignificant changes (P > 0.05) were observed considering the mean residence time of the extruded filaments within the extruder barrel, ranging from 235 to 246 s. HME and FDM-3DP processing parameters are presented in Table 5.

Table 5.

Hot-melt extrusion and fused deposition 3D printing process parameters.

Formulation	CIP-3DP-SR								CIP-3DP-IR
	Run 1	Run 2	Run 3	Run 4	Run 5	Run 6	Run 7	Run 8
HME temperature (°C)	140								90
Screw speed (rpm)	50
HME residence time (s)	243	239	241	230	239	246	241	236	235
FDM-3DP temperature (°C)	186
Printing speed (s/mm)	10
Printing time (s)	35	33	37	33	33	35	37	35	32

3.2. The Repka-Zhang test

Commercially available 3D printers are most widely used in the plastic industry. Therefore, we employed the most commonly used plastic-like filaments of PLA polymer as a reference to compare differences between commercial filaments and HME extruded filaments (Table 6), given that this filament exhibits appropriate stiffness, toughness, and melt viscosity characteristics (Zhang et al., 2017). Brittle filaments are easily crushed by feeding gears within the 3D printer, whereas soft filaments exhibit low abrasion resistance and cannot be conveyed by gears, thus blocking the head of the printer (Wang et al., 2020). Therefore, the breaking force, distance, and stress of the HME filament were evaluated and compared with previously published FDM-3DP literature. The breaking stress was calculated by dividing the breaking force by the cross-sectional area of the extruded HME filaments. The greater the breaking stress of both filaments, the greater the hindrance due to the filament texture. Similarly, the longer the breaking distance, the softer the filament (Wang et al., 2020). The mechanical characteristics of both filaments favored the FDM-3DP process, based on our earlier investigations (Wang et al., 2020).

Table 6.

The 3-point bend test results for HME Filaments (mean ± standard deviation, n = 5).

Filament	Force (g)	Stress (g/mm3)	Distance (mm)	Printability
CIP-3DP-IR	272.3 ± 11.5	120.0 ± 5.0	5.7 ± 0.2	Yes
CIP-3DP-SR	501.5 ± 14.4	221.1 ± 6.3	5.7 ± 0.1	Yes
PLA	1019.9 ± 27.6	449.6 ± 12.1	4.2 ± 0.1	Yes

CIP, ciprofloxacin; HME, hot-melt extrusion; IR, immediate release; PLA, polylactic acid; SR, sustained release.

3.3. FDM-3DP of ocular inserts

The capabilities of 3DP over conventional pharmaceutical manufacturing processes were exploited to design a novel complex-structured ocular insert. We employed thermoplastic HME extruded filaments as the FDM-3DP ink. The FDM-3DP inserts were fabricated by heating an extrusion nozzle to induce filament melting. The filament was converted into inserts by depositing the molten matrix in layers on the printing platform when it passed over the surface (Fig. 5). CIP-HCL ocular inserts were designed as flat oval objects. Based on factorial design variables, different layer heights, shell thicknesses, and infill densities were utilized to optimize the insert release profile. Thus, by harnessing established correlations between the input parameters (dimensions) and drug release kinetics of the inserts, the performance of formulated ocular inserts may be adjusted with simple mouse clicks (Vo et al., 2020). It is worth mentioning that the 3DP parameters for the immediate release insert (CIP-3DP-IR) were similar to those of the optimized design for the CIP-3DP-SR-OPT insert to perform an appropriate comparison.

Figure 5.

Layer-by-layer three-dimensional printing of the ocular insert using fused deposition modeling.

3.4. Statistical analysis of the applied experimental design

A three-factor model with a two-level full factorial experimental design was applied to examine all possible combinations of all factors with their respective levels in a fully randomized order. Table 3 presents the observed responses for the eight experimental runs created using the software. In the present study, a statistical model was applied to analyze the effect of selected independent variables on the in vitro release profiles of developed 3DP ocular inserts. Analysis of variance (ANOVA) was performed using relevant software to determine the significance of each model term, along with the interaction between different model terms and to calculate the adjusted coefficient of determination (R²), predicted R², the sum of squares, degree of freedom, and F-values. Contour plots, 3D response surface plots, cubes, and interaction graphs were generated to clarify the main effects of interaction on the responses. Numerical optimization was performed to achieve the study objectives with the target response (% release) values at the selected time points. Following data analysis using DesignExpert^® software, we observed that only two factors (X₁ and X₂) significantly affected both responses. The correlation data for fitting the selected factorial model of both responses are presented in Table 3. ANOVA testing revealed that the selected statistical model was significant (P < 0.05); thus, the selected model provided a good fit and could predict the response. The robustness of the statistical model was revealed by large F values, i.e., 118.71 for Y₁ and 146.17 for Y₂, indicating that the model was significant, with only a 0.01% chance that the large F-values could occur in response to noise. The predicted R² values, 0.9472 for Y₁ and 0.9570 for Y₂, were in reasonable agreement with the adjusted R² values (the difference was less than 0.2) of 0.9711 for Y₁ and 0.9765 for Y₂, suggesting a good predictive model. The obtained signal-to-noise ratios (adequate precision > 4), 24.299 for Y₁ and 27.324 for Y₂, indicated an adequate signal to navigate the design space. Box-Cox plots were analyzed for any recommended data transformations; however, no transformation was required for either response and no outliers were observed in the statistical model. The following regression model equations could predict each response in terms of coded factors:

$$\begin{gathered} {\mathbf{Y}}_1=47.88-{8.20}^{\ast }{\mathrm{X}}_1-{4.73}^{\ast }{\mathrm{X}}_2 \\ {\mathbf{Y}}_2=84.64-{5.64}^{\ast }{\mathrm{X}}_1-{8.46}^{\ast }{\mathrm{X}}_2 \end{gathered}$$

The intercepts (+47.88, Y₁ and + 84.64 for Y₂) represent the average responses of the eight experimental runs. In the regression equation above, a positive sign before each independent variable in the model equations reveals that the response increases with increasing factor level, whereas the opposite is true for the negative sign (Bolton and Bor, 2003). The coefficient values indicate the degree to which the independent variable contributes to the response. The two-dimensional contour plots, along with their respective 3D plots for both responses, are illustrated in Figs. 6-7.

Figure 6.

Two-dimensional contour (A&B) and three-dimensional response surface (C&D) plots for both responses

Figure 7.

Interaction graphs for both responses

3.5. Effect of the investigated 3DP parameters on drug release

The mean drug release (%) from the 3DP insert was significantly decreased with an increase in insert thickness (X₁) and infill density (X²), whereas layer height did not impact drug release, which could be due to the small difference between the two levels tested. Based on both regression equations, insert thickness negatively impacted both responses. Considering thin inserts, the edge effects are negligible owing to the high “surface area to insert thickness” ratio (Klose et al., 2008). Therefore, the drug release rate was higher from 0.3 mm-thick ocular inserts than from those with a thickness of 1.0 mm. 3DP inserts with high infill density (100%) exhibited considerably slower rates of CIP-HCL release than low infill density (75%) inserts. This observation could be attributed to the small pores or channels for water penetration and, thus, the relatively low surface area available for dissolution (Palekar et al., 2019; Wang et al., 2020).

3.6. Optimization and validation of the experimental design

Staphylococcus aureus, Coagulase-negative Staphylococci, Streptococcus pneumoniae, and P. aeruginosa are the predominant pathogens that cause ocular infections. The MIC 90 value of CIP is reportedly 64 μg/mL against fluoroquinolone-resistant S. aureus and 0.125 μg/mL against P. aeruginosa (Kowalski et al., 2003). Therefore, the prepared inserts should release at least 25% of their CIP content starting at the first time point (30 min) to supersede the highest required MIC 90 value (64 μg/mL) and afford superior therapeutic outcomes. Moreover, the main objective of the present study was to prepare sustained-release ocular inserts of CIP-HCL; therefore, the target drug release at 6 h was set between 60–75% (Table 8).

Table 8.

Criteria of independent and dependent variables for optimization of ciprofloxacin hydrochloride release from the 3DP ocular inserts.

Variable	Goal	Lower limit	Upper limit	Lower weight	Upper weight	Importance
X1 – insert thickness (mm)	equal to 1	0.3	1	1	1	+++
X2 – insert infill (%)	equal to 100	75	100	1	1	+++
X3 – layer height (mm)	within range	0.1	0.4	1	1	+++
Y1; Release at 0.5 h	within range	30	40	1	1	+++
Y2; Release at 6 h	within range	60	75	1	1	+++

3DP, three-dimensional printing.

The 3DP process was optimized by setting the goals of independent variables and responses and applying the global desirability function (D). Based on these criteria, a desirability plot was generated, with a D value of 1.0. To achieve the desired responses with 95% confidence intervals (CI), the software proposed a set of parameters similar to Run# 1 to fulfill the optimum formulation requirements. Based on these variables, their respective levels could result in a formulation with a % release of 32.5 and 70.1 at 0.5 (Y₁) and 6 h (Y₂) time points, respectively. The solution was graphically represented using the cube plots shown in Fig. 8. A validation trial was conducted in triplicate to compare the observed % release values against software-predicted values. The mean of the observed target release values was within 95% Cis of predicted values for the experimental model (Table 9).

Figure 8.

Cubes showing the desirability and predicted response for the software suggested solution based on the criteria of the optimization step.

Table 9.

Results of validation trials against the software suggested solution (Mean, n=3).

Response (%)	Predicted value	Results of validation trials	95% CI (Low) for mean	95% CI (High) for mean
Release at 0.5 h	34.9	32.5	32.2	37.6
Release at 6 h	70.4	70.1	67.6	73.0

CI, confidence intervals.

3.7. Characterization of CIP-3DP Inserts

3.7.1. Appearance, weight, uniformity of thickness, drug content, and surface pH

The molded inserts were evaluated for appearance, weight, uniformity of thickness, drug content, surface pH, moisture loss, absorption, and swelling index (Table 10). The weight of inserts for both formulations ranged between 5.0 and 5.3 mg. The ocular insert thickness was ~0.3 mm (Run# 2,4,6,8) and ~1.0 mm (Run# 1,3,5,7). An insert thickness of up to 1.0 mm allows easy topical ocular application and avoids foreign body sensation after topical application, thereby improving patient compliance. The drug content was acceptable (within ±10 % of label claim), varying between 91.1 and 103.7 %, with standard deviation (SD) values of < 5.0% for all runs, confirming the uniformity of drug content within the polymeric matrix owing to the molecular level mixing of the active ingredient, polymers, and plasticizers during HME processing (Alzahrani et al., 2022b). The surface pH of all inserts ranged from 7.3 to 7.4, which is favorable for ocular application because it is compatible with the tolerable pH range of the eye (3.0 to 8.6) and the normal human tear pH range (6.5 to 7.6) (Abelson et al., 1981).

Table 10.

Thickness, weight, drug content, surface pH, and moisture absorption and loss (%) of CIP-3DP-SR (8 runs) and CIP-3DP-IR inserts (mean ± SD, n = 3).

Formulation	CIP-3DP-SR								CIP-3DP-IR
	Run 1	Run 2	Run 3	Run 4	Run 5	Run 6	Run 7	Run 8
Thickness (μm)	996.7 ± 0.1	303.3 ± 0.1	1046.7 ± 0.1	310.0 ± 0.1	1043.3 ± 0.1	286.7 ± 0.1	1063.3 ± 0.1	1013.3 ± 0.4	991.9 ± 0.1
Weight (mg)	5.1 ± 0.1	5.2 ± 0.1	5.3 ± 0.1	5.1 ± 0.2	5.2 ± 0.1	5.1 ± 0.1	5.1 ± 0.1	5.1 ± 0.1	5.0 ± 0.2
Drug content (%)	95.4 ± 1.8	96.4 ± 3.3	93.2 ± 1.0	91.1 ± 2.5	91.9 ± 2.4	103.7 ± 1.6	96.0 ± 2.9	92.6 ± 1.59	93.4 ± 1.4
Surface pH	7.3 ± 0.2	7.3 ± 0.1	7.4 ± 0.2	7.3 ± 0.1	7.4 ± 0.1	7.3 ± 0.1	7.3 ± 0.1	7.4 ± 0.1	7.4 ± 0.1
Moisture absorption (%)	14.3 ± 0.8	14.0 ± 1.9	9.2 ± 1.4	11.3 ± 3.8	11.4 ± 3.0	10.9 ± 1.9	12.8 ± 0.6	13.2 ± 0.9	15.9 ± 1.3
Moisture loss (%)	5.3 ± 0.6	7.0 ± 1.1	6.5 ± 1.9	3.3 ± 1.0	5.4 ± 2.6	4.8 ± 1.2	2.3 ± 0.5	5.5 ± 1.9	7.2 ± 0.9
Physical appearance

CIP, ciprofloxacin; IR, immediate release; SR, sustained release.

3.7.2. Moisture loss and absorption and SI

The moisture absorption and loss (%) ranged from 9.2 ± 1.44 to 14.3 ± 0.79 % and 2.3 ± 0.54 to 7.0 ±2.10, respectively, for all runs. No changes in the insert integrity were detected based on the physical appearance examination, even under extremely humid and dry conditions.

SI demonstrates the water-retaining capability of polymers. The water absorption and swelling of the polymer increased with time, which was attributed to the polymer hydrophilicity. The SI of the formulation is crucial because it indicates the bioadhesive characteristics of inserts. The inserts did not break upon swelling and hydrated quickly, with the CIP-3DP-IR insert achieving hydration of approximately 32% in 30 min (Figure 9); however, the lead CIP-3DP-SR-OPT insert only reached 15% in 30 min. The markedly high SI value, which was observed with immediate-release ocular inserts containing mainly PEO N10, could be attributed to the high hydrophilicity of this polymer when compared with that of the HPC polymer, thus forming a viscous gel layer during swelling.

Figure 9.

Swelling index (%) of immediate (CIP-3DP-IR) and sustained (CIP-3DP-SR) release ciprofloxacin hydrochloride FDM-3DP ocular inserts in isotonic phosphate buffer saline (mean ± standard deviation [SD], n = 3).

3.7.3. DSC

Thermal curves of pure CIP-HCL, polymers (PEO N10 and HPC), plasticizers (sorbitol and PEG 4000), corresponding PMs, and drug-loaded inserts containing 6% w/w of the drug are shown in Fig. 10. Pure CIP-HCL exhibited a sharp endothermic peak at 161°C due to the crystallinity of the drug. In contrast, PEO and HPC exhibited endothermic melting peaks at 69and 90°C, respectively, demonstrating their melting transitions. The thermograms of both polymers revealed that they could soften at ≥ 90°C. PEG 4000 and sorbitol DSC thermograms showed endothermic melting peaks at 60 and 100°C, respectively. The absence of an endothermic drug peak in the one-phase solid solutions of the extruded filament and CIP-FDM-3DP insert DSC curves indicated that the drug was converted to its amorphous form within the polymer matrix.

Figure 10.

DSC thermograms of pure active and inactive ingredients, physical mixtures, and ciprofloxacin hydrochloride FDM-3DP ocular inserts. CIP, ciprofloxacin; DSC, differential scanning calorimetry; FDM-3DP, fused deposition modeling three-dimensional printing; IR, immediate release; SR, sustained release.

3.7.4. FTIR

The FTIR spectra of pure CIP-HCL, PEO N10, HPC JXF, PEG 4000, sorbitol, HME extruded filaments, PMs, and CIP-FDM-3DP inserts were examined to determine the incompatibility between active and inactive ingredients in the formulated ocular inserts, as shown in Fig. 11. The FTIR spectra of the pure drug exhibited a carbonyl group stretching band at 1722 cm⁻¹, stretching vibration of the C─F bond at 1290 cm⁻¹, and C–H stretching vibration band of the phenyl ring at 3043 and 2918 cm⁻¹ (Youssef et al., 2021). The FTIR spectra of PM, extruded filaments, and inserts of CIP-3DP-IR formulation showed high similarity to the pure PEO N10 spectrum, given that they primarily contained this polymer (69.375% w/w). In contrast, FTIR spectra of PM, extruded filaments, and inserts of the CIP-3DP-SR-OPT formulation exhibited several characteristic peaks of the pure HPC spectrum, given that this polymer represents the main ingredient of the formulation (91% w/w). All characteristic peaks were present in both drug-loaded inserts and PMs, with no new peaks interfering with the FTIR spectra. Therefore, our findings suggest no physical or chemical incompatibilities between the active and inactive ingredients of either formulation.

Figure 11.

FTIR spectra for (A) Ciprofloxacin hydrochloride, (B) HPC JXF, (C) PEO N10, (D) PEG 4000, (E) Sorbitol, (F) PM of CIP-3DP-IR, (G) Extruded filaments of CIP-3DP-IR, (H) CIP-3DP-IR inserts, (I) PM of CIP-3DP-SR insert, (J) Extruded filaments of CIP-3DP-SR, and (K) CIP-3DP-IR inserts.

3.7.5. SEM

SEM micrographs of the pure CIP-HCL powder revealed needle-like crystals forming crystalline aggregates (Fig. 12). In addition, CIP-HCL retained its crystalline nature in the PM of the formulation. The HME filament and 3DP inserts of the optimized CIP-3DP-SR-OPT formulation showed a smooth surface. Therefore, the SEM micrographs complemented the DSC results, showing an amorphous drug within the polymeric matrix.

Figure 12.

SEM micrographs of ciprofloxacin hydrochloride powder (100× (a.1) and 1000× (a.2) magnification), physical mixture (100× (b.1) and 1000× (b.2) magnification), extruded filaments (500× (c.1), 1000× (c.2), and 10,000× (c.3), magnification), and optimized design of CIP-3DP-SR inserts with 1.0 mm thickness and 100% infill density (25× (d.1), 100× (d.2), and 500× (d.3)). Abbreviations: 3DP, three-dimensional printing; CIP, ciprofloxacin; FTIR, Fourier transform infrared; SEM, scanning electron microscopy; SR, sustained release.

3.7.6. In vitro release

The in vitro release profiles of CIP from CIP-3DP-SR (eight runs) and CIP-3DP-IR inserts are presented in Fig. 13 A-B. Drug release significantly decreased with an increase in insert thickness (Fig. 13C) and infill density (Fig. 13D). Considering thin inserts, the edge effects are negligible owing to the high “surface area to insert thickness” ratio (Klose et al., 2008). Therefore, the drug release rate was higher from 0.3 mm-thick ocular inserts than from 1.0 mm-thick inserts. 3DP inserts with high infill density (100%) exhibited much slower rates of CIP-HCL release than low infill density (75%) inserts. This observation could be attributed to the small pores or channels for water penetration, thereby a relatively low surface area available for dissolution (Palekar et al., 2019; Wang et al., 2020).

Figure 13.

(A) In vitro release profiles of ciprofloxacin hydrochloride from CIP-3DP-SR inserts (8 runs), (B) In vitro release profile of ciprofloxacin hydrochloride from CIP-3DP-IR vs. the optimized design of CIP-3DP-SR inserts, (C) Effect of insert thickness on CIP-HCL release from CIP-3DP-SR inserts, and (D) Effect of infill density on CIP-HCL release from CIP-3DP-SR inserts (n=3). Abbreviations: CIP, ciprofloxacin; IR, immediate release; SR, sustained release.

To elucidate the mechanism of drug release, the release data for the optimized CIP-3DP-SR-OPT inserts were fitted to five main conventional release models: zero-order (R² = 0.6305), first-order (R² = 0.8167), Higuchi (R² = 0.9849), Hixson-Crowell (R² = 0.7619), and Korsmeyer-Peppas (R² = 0.9996). Mathematical model fitting revealed that the release profile of the ocular insert followed the Korsmeyer-Peppas model (highest R²). The calculated slope value (n = 0.32) from the selected model (n < 0.5) indicated a Fickian drug release profile from the polymeric matrix governed by a diffusion mechanism. To confirm the control release mechanism, the amount released was plotted versus the square root of time (Higuchi model). The plot is a straight line with a high R² value of 0.9849.

3.7.7. Evaluation of antibacterial activity

The objective of the present study was to clarify the correlation between the release profile and antimicrobial activity of the optimized design of the CIP-3DP-SR-OPT insert (Fig. 14). No bacterial growth was observed in any withdrawn sample from the receiver medium (0.5 to 24 h), indicating that the CIP concentration in the receiver medium exceeded the MIC 50 value (1.25 μg/mL) against P. aeruginosa across all study time points. Notably, released samples were diluted 10-fold before evaluating the antibacterial activity of the CIP-3DP-SR ocular inserts. Thus, an MIC of 50 μg/mL was achieved at any time before the first release time point (30 min). Based on the strong correlation revealed in the present study, the FDM-3DP inserts could afford superior therapeutic outcomes.

Figure 14.

Relationship between the release profile of ciprofloxacin hydrochloride from the optimized design of the CIP-3DP-SR ocular insert and the antimicrobial activity against Pseudomonas aeruginosa (% growth and release against time). Abbreviations: CIP, ciprofloxacin; IR, immediate release; SR, sustained release.

3.7.8. Ex vivo transcorneal permeation

The transcorneal flux of CIP across isolated rabbit corneas (Fig.15) from the commercial eye drop solution, CIP-3DP-IR, and the optimized design of the CIP-3DP-SR-OPT insert are shown in Fig. 16. The flux from eyedrops (0.33± 0.03 μg/ min/cm²) was higher than that of the optimized design of CIP-3DP-SR insert (0.18 ± 0.01 μg/min/cm²). This finding is consistent with previous reports (Thakkar et al., 2021), wherein the commercial eye drop solution showed a higher flux than the corresponding moxifloxacin HME ocular inserts. These observations could be attributed to the presence of CIP in solution form (eye drops), while the drug is present in solid form in the ocular inserts, thereby resulting in lower flux. Therefore, the CIP-3DP-IR insert was included as a control for appropriate comparison with the CIP-3DP-SR-OPT insert.

Figure 15.

A schematic graph of the ex-vivo drug transcorneal permeation from optimized ciprofloxacin hydrochloride FDM-3D printed inserts was studied using Franz diffusion cells. 3D, three-dimensional; FDM, fused deposition modeling.

Figure 16.

Transcorneal ciprofloxacin hydrochloride flux (μg/min/cm²) from commercial solution eyedrops (CIP-C), CIP-3DP-IR, and CIP-3DP-SR inserts (Mean ± standard deviation [SD], n = 3). Abbreviations: 3DP, three-dimensional printing; CIP, ciprofloxacin; IR, immediate release; SR, sustained release.

The CIP-3DP-IR insert achieved an approximately 1.4-fold increase in drug flux (0.25 ± 0.02 μg/min/cm²) when compared with the CIP-3DP-SR-OPT insert. This observation could be due to the composition of the CIP-3DP-IR insert, mainly comprising the immediate-release PEO N10 polymer, which lacks the controlled release behavior of the HPC polymer. Therefore, the drug can diffuse out of the polymeric matrix at a faster rate, affording enhanced flux. Importantly, our findings revealed that CIP-3DP-SR-OPT inserts achieved a CIP concentration of 2.9 ± 0.5 μg/ mL in the receiver compartment at 30 min, exceeding the MIC 50 (1.25 μg/mL) against P. aeruginosa.

3.7.9. Ex vivo evaluation of mucoadhesive strength

Mucoadhesive dosage forms enable prolonged residence on the ocular surface and improve ocular permeation, resulting in better therapeutic outcomes (Shaikh et al., 2011). The interaction of a polymer with the mucin coat covering the conjunctiva and corneal surfaces of the eye facilitates ocular mucoadhesion (Qi et al., 2007). In polymers, carboxyl, hydroxyl, and sulfate groups can form electrostatic, hydrophobic, and hydrogen-bonding interactions with the underlying surface. Hydrogen bonding appears to be the most significant noncovalent force (Qi et al., 2007)(Robinson and Mlynek, 1995). Precorneal retention of the active pharmaceutical ingredient may be enhanced by HPC, an anionic polymer containing carboxylic groups with good adhesive qualities, which may delay washing (Nichols et al., 1985). Consequently, the mucoadhesive capabilities of CIP-3DP-IR and CIP-3DP-SR-OPT inserts were evaluated using freshly excised rabbit corneas, as demonstrated in Fig. 17. As the testing time increased for both inserts, the mucoadhesive force and WOA increased concurrently; the immediate-release inserts exhibited the highest mucoadhesive force and WOA values. Furthermore, the two mucoadhesive properties were intimately linked: the more significant the mucoadhesive force, the higher the WOA for the ocular inserts. Bioadhesive and mucoadhesive properties have been attributed to diverse theories, including electronic, diffusion, adsorption, wetting, and interpenetration capacities (Ludwig, 2005). For the polymer included in the drug delivery system to function as an effective mucoadhesive agent, it must be in perfect contact with the mucin layer enveloping the corneal surface (Ludwig, 2005). The polymer chains must be mobile and flexible to diffuse into the mucus and subsequently penetrate to a depth that allows entanglement with mucin, forming a robust network. In addition, interactions between polymers and mucins include hydrogen bonding, electrostatic attraction, and hydrophobic repulsion (Ludwig, 2005). However, a specific chain length is necessary to achieve the desired interpenetration and molecular entanglement between the polymer and mucus layer. The minimum molecular weight necessary for effective mucoadhesion is 100 kDa. (Ludwig, 2005). Polymers with molecular weights exceeding 100 kDa may exhibit excessive cross-linking between polymer chains, reducing the chain length available for interfacial penetration and tangling with the mucus layer (Madsen et al., 1998). The superior adhesive properties of PEO N10 may be attributed to the lower molecular weight (10 kDa) of PEO polymer JXF than that of the HPC polymer (140 KDa).

Figure 17.

Mucoadhesive forces (A) and work of adhesion (B) for immediate and sustained release of FDM-3DP ocular ciprofloxacin hydrochloride inserts (means ± standard deviation [SD], n =3). Abbreviations: FDM-3DP, fused deposition modeling three-dimensional printing.

3.7.10. Stability

The physical and chemical stabilities of the CIP-3DP-SR-OPT inserts were evaluated over 3 months of storage at room temperature (last time point analyzed). The effect of storage conditions on the CIP-HCL content and release profile is shown in Fig. 18, and the effect on thermal behavior is shown in Fig. 10. Some statistically insignificant changes (P > 0.05) were observed in the CIP-HCL content and drug release rate over the testing period. The three-month DSC thermogram of the fabricated inserts did not exhibit any CIP-HCL crystal-form transformation upon storage. Furthermore, no physical changes such as swelling or discoloration were detected.

Figure 18.

A) In vitro release profiles and B) CIP-HCL content of the CIP-3DP-SR inserts over a three-month storage period at 25°C (mean ± standard deviation [SD], n = 3). Abbreviations: 3DP, three-dimensional printing; CIP, ciprofloxacin; IR, immediate release; SR, sustained release.

Conclusions

3DP is a major development in the manufacture of pharmaceutical dosage forms because it enables the printing of complex and customized dosage forms for immediate consumption based on the unique demands of the patient. Coupling HME with FDM-3DP technology was successfully utilized to fabricate customized antibacterial ocular CIP-HCL inserts. The in vitro and ex vivo performances of the prepared 3DP inserts were examined. The DOE approach was applied to optimize the inserts and achieve a sustained release profile for at least 24 h. The optimized design afforded excellent stability at room temperature over three months, with no significant change in drug content, release profile, or thermal behavior. In addition, the optimized design demonstrated an amorphous solid dispersion, no physical or chemical incompatibility between the different ingredients, excellent mucoadhesive strength, excellent antibacterial activity, and prolonged transcorneal drug flux when compared with commercial eye drops and immediate-release inserts. The flexible printed inserts allow bending and placement of inserts to achieve the desired dosages of CIP-HCL while ensuring patient compliance. For ocular delivery of CIP-HCL, these FDM-3DP polymeric inserts could afford a longer-lasting effect than solution/topical eye drops. However, future studies, including in vivo ocular irritation, efficacy, and ocular biodistribution studies, are required for further development. In conclusion, FDM-3DP inserts could be a promising drug delivery platform for CIP in managing BK and other ocular bacterial infections.

Table 7.

Results obtained from ANOVA testing and fit statistics for the two-level randomized factorial design during the optimization of ciprofloxacin hydrochloride release from the 3DP ocular inserts.

Source	Release at 0.5 h (Y1, %)				Release at 6.0 h (Y2, %)
	Sum of squares	Degree of freedom	F-value	P-value@	Sum of squares	Degree of freedom	F-value	P-value
Model	716.52	2	118.71	< 0.0001	851.10	2	146.17	< 0.0001
X1	537.92	1	178.24	< 0.0001	254.25	1	87.33	0.0002
X2	178.61	1	59.18	0.0006	596.85	1	205.02	< 0.0001
Residuals	15.09	5	3.02		14.56	5	2.91
Total	731.61	7			865.66	7
Adjusted R2*	0.9711				0.9765
Predicted R2	0.9472				0.9570
Adequate precision#	24.299				27.324

^@P-values < 0.05 indicate significant model terms, and values >0.1000 indicate the model terms are not significant.

^*The difference between Adjusted R2 and Predicted R2 should be < 0.2

^#Adequate precision measures the signal-to-noise ratio. A ratio of > 4 is desirable.

3DP, three-dimensional printing.

Abbreviations:

ATCC

American Type Culture Collection

BK

Bacterial keratitis

CLS

Clinical and Laboratory Standards Institute

DoE

Design of experiment

FDM-3DP

Fused deposition modeling three-dimensional printing

HME

Hot-melt extrusion

IPBS

Isotonic phosphate-buffer saline

MIC

Minimum inhibitory concentration

PM

Physical mixture

STF

Simulated tear fluid

3DP

Three-dimensional printing

WOA

Work of adhesion

SR

Sustained release

IR

Immediate release

OPT

Optimized