Microneedle-mediated intrascleral delivery of in situ forming thermoresponsive implants for sustained ocular drug delivery

Abstract

Objectives

This paper describes use of minimally invasive hollow microneedle (HMN) to deliver in situ forming thermoresponsive poloxamer-based implants into the scleral tissue to provide sustained drug delivery.

Methods

In situ forming poloxamer formulations were prepared and investigated for their rheological properties. HMN devices 400, 500 and 600 μm in height were fabricated from hypodermic needles (i.e. 27, 29 and 30 G) and tested for depth of penetration into rabbit sclera. Maximum force and work required to expel different volumes of poloxamer formulations was also investigated. Release of fluorescein sodium (FS) from intrasclerally injected implants was also investigated. Optical coherence tomography (OCT) was used to examine implant localisation and scleral pore-closure.

Key findings

Poloxamer formulations showed Newtonian behaviour at 20°C and pseudoplastic (shear-thinning) behaviour at 37°C. Maximum force and work required to expel different volumes of poloxamer formulations with different needles ranged from 0.158 to 2.021 N and 0.173 to 6.000 N, respectively. OCT showed intrascleral localisation of implants and scleral pore-closure occurred within 2–3 h. Sustain release of FS was noticed over 24 h and varied with depth of implant delivery.

Conclusions

This study shows that the minimally invasive HMN device can localise in situ forming implants in the scleral tissue and provide sustained drug delivery.

Introduction

Sight threatening eye diseases such as age-related macular degeneration, diabetic retinopathy, diabetic macular oedema and uveitis originate in the posterior segment of the eye. These diseases are chronic in nature, and current practice includes frequent intravitreal injections of therapeutic agents (e.g. flucinolone acetonide, ranibizumab). Even though intravitreal injections provide direct delivery of therapeutic agents into the eye, this method is invasive and associated with severe side effects, such as high therapeutic dosage-induced ocular toxicity, pain and discomfort, vitreous haemorrhage, retinal detachment, endophthalmitis, and cataract development.^[1–3] On the other hand, the periocular (or transscleral) route that includes retrobulbar, peribulbar, subtenon and subconjunctival routes is considered to be less painful and most efficient route of drug delivery to the posterior segment of the eye that poses reduced risk of endophthalmitis and retinal damage.^[4]

The human sclera (the white, structural sheath around the circumference of the eye) covers 95% of the eye's surface area and equates to approximately 17 cm². It is a largely acellular tissue composed of interweaving collagen fibres embedded in an aqueous, glycosaminoglycan matrix.^[5] The permeability of the sclera is comparable with the corneal stroma, and drug lipophilicity does not appear to affect scleral permeability markedly.^[6] The transscleral or periocular route of drug delivery has shown to permit large range of molecules including corticosteroids,^[7,8] antisense oligonucleotides,^[9] immunoglobulins^[10] and DNA^[11] to the posterior segment of the eye. For example, rabbit sclera is permeable to drugs up to a MW of 150 kDa, but permeability declines exponentially with increasing MW and molecular radius.^[12] However, transscleral delivery is associated with low bioavailability of drug in the vitreous, which is attributed to loss of drug from the periocular space (via reflux), presence of the blood retinal barrier, choroidal circulation currents, binding of drugs to tissue proteins, presence of efflux transporters and prolonged time for diffusion of drugs across sclera (e.g. transscleral route).^[13]

To overcome transscleral permeation of drug molecules, a number of studies have shown surgical administration of drug-loaded solid implants within the scleral tissue (i.e. intrascleral delivery) that can provide sustained posterior drug delivery.^[14–16] However, surgical method of administration is highly invasive and causes higher tissue trauma, taking longer healing time, retinal detachment and increased costs after implantation, such as the need for anti-inflammatory eye drops to avoid any chances of post-surgical infections. Furthermore, precise localisation of implants within the scleral tissue is not possible by using traditional hypodermic needles or by surgical interventions because of relatively thin scleral tissue.

Microneedles that are third-generation minimally invasive devices have been developed to enhance transdermal drug delivery for a wide range of therapeutic molecules.^[17] Microneedles are typically 25–2000 μm in height and have been fabricated from a wide range of materials and in different shapes. Microneedles can be fabricated either into hollow or solid needles. To date, limited work has been done in the use of microneedles for ocular drug delivery applications. Using hollow microneedle (HMN) allows a minimally invasive means of ocular drug delivery, where the drug delivery system can be precisely localised within ocular tissues such as in the suprachoroidal space, sclera tissue or other ocular tissues.^[18,19] However, to date, no studies have been reported on the use of microneedles to inject in situ implant-forming drug delivery systems into the scleral tissue, which could potentially provide a minimally invasive means of sustained intrascleral drug delivery.

Therefore, in this study, for the first time, we have investigated the application of HMNs to precisely deliver a thermoresponsive-based in situ implant-forming poloxomer formulations into the scleral tissue. The poloxomer formulations were engineered to be free flowing at ambient conditions to allow administering using HMNs but form an implant at the physiological temperature of the eye and provide sustained delivery. In so doing, we have studied the rheological properties of the poloxamer formulations with and without model drug (i.e. fluorescein sodium, FS). Fabricated HMN devices from four different types of hypodermic needles with different microneedle heights. Investigated forces required for syringeability of poloxamer solutions through HMN devices. HMN device was used to administer the implant-forming FS-loaded poloxamer solutions within the rabbit scleral tissue, and the release was examined overtime. Furthermore, for the first time, in this study, we have used optical coherence tomography (OCT) to visualise the implant formation in the sclera and also to characterise the scleral pore-closure following HMN application.

Materials and Methods

Materials

FS (MW 376.27 Da) was purchased from Sigma-Aldrich, Dorset, UK, and Pluronic® F-127 (poloxamer 407) (MW 12 500 Da) and Pluronic® F-87 (poloxamer 237) (MW 7700 Da) were purchased from BASF Chemical Company, Ludwigshafen, Germany. Hypodermic needles 26 G (Terumo Neolus®, 0.45 × 12 mm), 29 G (Terumo Myjector® 1.0 ml, 0.33 × 12 mm) and 30 G (BD Micro-Fine™ 0.3 ml, 0.3 × 8 mm) needles were used as supplied. A 1 ml Terumo (Terumo Medical Corporation, Murray Hill, NJ, USA) syringes were used as supplied.

Preparation of poloxamer solutions

A series of thermoresponsive poloxamer formulations were prepared using the cold method, as shown in Tables 1 and 2. Distilled water was cooled to 4°C, then poloxamers 407 and 237 were slowly added to the distilled water in the required proportions with continuous agitation. The polymeric dispersions were then left at 4°C until a clear solution was obtained.^[20]Table 2 represents poloxamer formulations F1–F6 containing 0.5% w/w of FS were prepared as discussed earlier.

Table 1.

Gelation temperatures of poloxamer mixtures, n = 3

Poloxamer	Concentration (% w/w)	Mean gelation temperature (±standard deviation) (°C)
407	10	>40
	15	>40
	20	23.9 ± 0.2
	25	15.9 ± 0.3
	30	11.7 ± 0.2
237	15	>40
	20	>40
	25	>40
	30	39.6 ± 0.3
407/237	10/2.5	>40
	10/5	>40
	10/10	>40
	10/15	>40
	10/20	>40
	10/25	29.3 ± 0.2
	12/2.5	>40
	12/5	>40
	12/10	39.0 ± 0.2
	12/15	34.3 ± 0.3
	12/20	22.5 ± 0.2
	15/2.5	36.3 ± 0.2
	15/5	34.8 ± 0.3
	15/10	17.3 ± 0.4

Table 2.

Effect of 0.5% w/w fluorescein sodium (FS) loading on gelation temperatures of selected poloxamer mixtures containing 12% w/w poloxamer 407 and variable concentrations of poloxamer 237, n = 3

Formulation	Poloxamer 237 (% w/w)	Mean gelation temperature (±standard deviation) (in °C)
		Concentration of FS (% w/w)
		0	0.5
F1	15	34.3 ± 0.3	30.7 ± 0.2
F2	16	30.4 ± 0.2	27.3 ± 0.2
F3	17	27.1 ± 0.1	25.6 ± 0.2
F4	18	24.8 ± 0.2	24.1 ± 0.3
F5	19	23.1 ± 0.3	22.2 ± 0.3
F6	20	22.5 ± 0.2	20.8 ± 0.6

Fabrication of microneedle adapters

Three microneedle adapters were fabricated by removal of the syringe head of 16 G intravenous catheters (BD Angiocath™, BD, Franklin Lakes, NJ, USA) and subsequently machined of the syringe head (Figure 1). The adapters were attached to the head of needles and imaged with a digital light microscope (GXMGE-5 digital microscope, Laboratory Analysis Ltd, Devon, UK). The microneedle adapters were machined to give a protrusion distance (i.e. needle heights) of 400, 500 and 600 μm, respectively, from the syringe head. Microneedle heights were measured using the ruler function of the microscope software.

Figure 1.

Schematic representation of (a) adapter fabrication from intravenous catheters and (b) attaching the adapter to hypodermic needle and syringe assembly to create hollow microneedles of desired heights.

Rheological studies of poloxamer solutions

Measurement of sol-gel transition temperatures of poloxamer solutions

Sol-gel transition temperatures of the poloxamer formulations were determined using an AR 2000 rheometer (T.A. Instruments, Surrey, UK) in flow mode with 4 cm diameter stainless steel parallel plate geometry. The formulations were heated between 5 and 40°C at a rate of 1°C/min under a constant shear rate of 5/s. Sol-gel transition temperatures were determined by plotting temperature as a function of viscosity (η), and the transition point was taken to be where the viscosity was halfway between the values for the solution and gel.^[20] Measurements were performed in triplicate.

Rheological characterisation at 20 and 37°C

The rheological behaviour of F1–F6 formulations at 20 ± 0.1°C and 37 ± 0.1°C was investigated using an AR 2000 rheometer in flow mode using a 4 cm diameter stainless steel parallel plate geometry. Formulations under a constant force of 5.0 N were exposed to a continuous shear rate ranging from 0 to 50/s. Viscosity measurements for each formulation were performed in triplicates.

Determination of maximum force and work required for injecting poloxamer formulations through hollow microneedle device

A Texture Analyser (TA-TX2, Stable Micro Systems, Surrey, UK), set in compression mode, was used to simulate the maximum forces and work (i.e. syringeability) required to inject the poloxamer formulations into scleral tissue at 20°C. The HMN devices tested for injections were consisted of three different types of needles, namely 26, 29 and 30 G, as these are commonly used for intravitreal injection. Initially, the maximum force and work required to expel different volumes, for example, 30, 50 and 100 μl of air, were determined as these represent the frictional component of the force and work required to inject the poloxamer formulations from the HMN device. A concentric cylinder probe (diameter 40 mm and depth 150 mm) was used to apply a force to the plunger of a syringe at a rate of 2 mm/s. Then the same process was carried out for syringes containing 30, 50 and 100 μl of F1–F6 formulations. Four replicate measurements were made in each case. The force and work were derived from the force-time plots, produced by Stable Micro Systems' Exponent software (v6.0.2.0, Stable Micro Systems). The maximum force and work were determined by measuring the peak maxima and the area under curve values, respectively, from the force-time plots. To calculate the maximum force and work directly attributable to the poloxamer formulations, the maximum force and work values for air (i.e. without formulations) were subtracted from the values obtained with formulations.

Measurement of penetration force of HMN into rabbit sclera

Rabbit eyeballs were obtained a day before the experiment from a local abattoir and were stored at −20°C. Before dissection, the rabbit eyeballs were defrosted in distilled water at 37°C for 6 h. Then, using a scapula, the intraocular and extraocular muscles, and connective tissue attached to the sclera of the rabbits' eyes were carefully removed. The thickness of the sclera in three areas, namely anterior, equator and posterior, were measured by using OCT (high-resolution EX1301 OCT Microscope, Michelson Diagnostics Ltd, Kent, UK) (Figure 2). The swept-source Fourier domain OCT system has a laser centre wavelength of 1305.0 ± 15.0 nm, facilitating real-time high-resolution imaging of the scleral and other ocular tissue layers (7.5 μm lateral and 10.0 μm vertical resolution). The sclera was scanned at a rate of up to 15 B-scans (two-dimensional (2D) cross-sectional scans) per second (scan width = 2.0 mm). 2D images were analysed using the imaging software ImageJ® (National Institutes of Health; http://rsbweb.nih.gov/ij/list.html). The scale of the image files obtained was 1 pixel = 4.2 μm. From this, the recorded average equatorial, anterior and posterior scleral thickness was 422 ± 57, 531 ± 82 and 675 ± 47 μm, respectively. Five measurements were made for each scleral position.

Figure 2.

Representative optical coherence tomography images of scleral tissue at different positions of the eye (a) equator, lateral to the middle of the lens; (b) anterior, at the edge of the orbit; and (c) posterior, lateral to the optic nerve. Scale bar = 400 μm.

It is important to note that only eyes that had sufficiently thick sclera of at least 400 μm at the equator, at least 500 μm at the anterior and at least 600 μm at the posterior were used for penetrative forces studies to prevent full puncture of the sclera and give erroneous results. Following thickness measurement an incision made behind the lens allowed for drainage of the vitreous humour and a 9 mm biopsy punch was used to punch out the scleral tissue at the three respective positions. The tissue was kept in distilled water at 37°C for a maximum of 1 h before use.

A Texture Analyser set to compression mode was used to determine the scleral penetration forces of the HMN (Figure 3). The syringe barrel of the HMN were attached to a concentric cylinder probe using cyanoacrylate adhesive, and the scleral tissue was placed onto the surface of a polystyrene hemisphere to mimic the curvature of the eye. It was made sure that the tips of the HMN were just above the surface of the sclera but did not contact with scleral tissue. Then, the HMN penetrated the sclera tissue to a depth of 400 μm for equatorial sclera, 500 μm for anterior sclera and 600 μm for posterior sclera at a rate of 0.1 mm/s. Five measurements were made for each scleral position. The penetration force was derived as the maximum force from the force-time plots.

Figure 3.

Schematic representation of Texture Analyser setup for determination of penetration forces of hollow microneedle into rabbit's scleral tissue.

Visualisation of implant formation using optical coherence tomography

In this study, whole rabbit eyeballs were defrosted in distilled water at 37°C for 6 h. The eyes where then placed into a small water-jacketed cylindrical container (diameter 18 mm) so that the rabbit eyes could be kept stationary at 37°C. For visualising the implant formation, a 50 μl of F6 formulation was injected into the sclera using the 400, 500 and 600 μm HMN device fitted with 30 G needles into the equatorial, anterior and posterior regions of the scleral tissue, respectively. The injection sites and injected formulation were visualised using OCT to confirm that the HMNs penetrated to the required depth and that formulation had been successfully injected and had formed a gel implant at 37°C (body temperature). High-resolution OCT images were taken at regular intervals to assess implant behaviour within the sclera and recovery of the scleral tissue at the injection site. To allow differentiation between the implants and different ocular tissues, false colours were applied using Ability Photopaint® Version 4.14 (Ability Plus Software Ltd, Crawley, UK). The experiment was setup so that the implant could be visualised at 37°C in the cylindrical container, thereby negating the chance of the poloxamer gel implant from turning back into a liquid state.

Ex vivo release studies through rabbit sclera

As detailed in Section ‘Measurement of penetration force of needles into rabbit sclera’, a 9 mm diameter scleral tissue samples were excised from the three positions (equatorial, anterior and posterior) of freshly defrosted rabbits' eyes and then positioned onto the receptor compartment of 5 ml Franz Diffusion Cells (PermeGear, Hellertown, PA, USA) such that there was intimate contact with the receptor fluid, 0.1 M phosphate buffered saline (PBS) at pH 7.4, which was thermostatically controlled at 37°C by a surrounding water jacket and stirred continuously at 600 rpm by a magnetic stirrer, as shown in Figure 4. Upturned 5 ml Franz-cell donor compartment lids were then used in conjunction with stainless steel clamps supplied with the 5 ml Franz-cells to hold the setup in place. Finally, a glass lid was placed on top to prevent desiccation of the scleral tissue. Volumes of 50 μl of formulations were injected into the scleral tissue to a depth of 400 μm for equatorial sclera, 500 μm for anterior sclera and 600 μm for posterior sclera using HMN device equipped with the appropriate adapter; a slight retraction of needle was done before injection of formulation to create a void space before applying pressure on the plunger to aid implant delivery in the sclera. Samples of 300 μl were then removed at defined time periods from the receptor compartments and replaced with 300 μl of fresh 0.1 M pH 7.4 PBS. The samples were then analysed for FS content using UV spectroscopy (PowerWave XS, Bio-Tek Instruments, Inc., Winooski, VT, USA) at 450 nm.

Figure 4.

Schematic representation of Franz cell setup used for ex vivo FS release studies.

Statistical analysis

The effects of the various experimental parameters on the experimental outcomes were performed with a one-way analysis of variance, where P < 0.05 was taken to represent a statistically significant difference. When there was a statistically significant difference, post-hoc Tukey's HSD tests were used. Drug release data were analysed using paired t-tests. In all cases, P < 0.05 was used to denote statistical significance.

Results and Discussion

In situ implant-forming systems are those that exist as a solution (i.e. liquid) under ambient conditions yet undergo a phase transition to a gel when exposed to a physiological environment or external stimulus. Factors that can trigger implant formation include a change in temperature (as in the case of poloxamers), pH, the presence of ions (usually divalent ions, e.g. Mg²⁺, Ca²⁺), phase inversion (unfavourable mixing of polar and organic phases), enzymatic and photo-cross-linking.^[21]

Several commercial products for anterior ocular use based upon in situ gelling systems exist, showing the viability of such systems. These include Timoptic-XE® (Merck & Co., Whitehouse Station, NJ, USA) containing gellan gum which has a temperature- and cation-dependent mechanism of gel formation^[22] and Virgan® (Spectrum Thea Pharmaceuticals Ltd, Macclesfield, Cheshire, UK) that contains carbomer 974 that has both temperature- and pH-dependent mechanism of gel formation.^[23]

In this study, triblock co-polymers (poly(ethylene oxide)-b-poly(propylene oxide)-b-poly(ethylene oxide)), commercially called as poloxamers, were used as thermoresponsive-based in situ implant-forming drug delivery system. poloxamers are fluid below a characteristic gelation temperature but convert to a gel above this temperature.^[24] At low concentrations of poloxamer, in cold-water, solvation of the polymer chains occurs. With increasing temperature, desolvation of the hydrophilic chains occurs because of hydrogen bond breakage, and hydrophobic interactions between PPO domains dominate, thus results in gel formation.^[25] The gel is micellar in nature and the liquid micellar phase, which is stable at low temperatures, converts to a cubic structure as temperature increases.^[25] Another form consisting of hexagonally packed cylinders occurs at even higher temperatures. Gel formation occurs only when poloxamer concentration is above the critical micellar concentration.^[26] Because of their good solubilising capacity, opacity and low toxicity, they have been used in ocular delivery of number of drugs to the anterior eye including pilocarpine,^[27] tropacamide,^[28] fluconazole^[30] and antibiotics.^[30,31] To the author's knowledge, this is the first time poloxamers have been injected intrasclerally to form an implant and assessed for sustained drug delivery.

High concentrations of poloxamers are usually needed (e.g. 20–30% w/w) to induce gelation at body temperature and the gelation itself occurs over a very narrow temperature range. To overcome this rapid transition and to improve mechanical/rheological properties, a mixture of two poloxamers with different MWs and relative hydrophilicities/hydrophobicities may be used.^[21] Ideally, a gelation temperature of 20–37°C ensures complete gelation when injected into the eye, that is, at 37°C, yet it is free-flowing and easily injected at room temperature, that is, at 20°C. In this study, two poloxamer types were used to this affect namely poloxamer 407 and poloxamer 237.

The gelation temperature of 10%, 15%, 25% and 30% w/w of poloxamer 407 gels was >40, >40, 15.9 and 11.7°C, respectively (Table 1). However, a concentration of 20% w/w produced a mean gelation temperature of 23.9°C that was within the desired range, that is, 20–37°C. Poloxamer 237 formulations did not form a gel within the desired range at all concentration studied. On the other hand, different poloxamer mixtures containing 12% w/w of poloxamer 407 and variable concentrations of poloxamer 237 (15 and 20% w/w) showed gelling temperatures ranging from 22.5 to 34.3°C and were chosen for further investigations. A series of formulations were prepared that contains 12% w/w of poloxamer 407 and a range of poloxamer 237 (15–20% w/w) loaded with 0.5% w/w FS; these formulations were denoted as F1–F6, as shown in Table 2. It was seen that addition of FS resulted in the reduction of the gelation temperatures (Table 2). This may have been to ionic bond, hydrogen bond or salting-out effect, which could be due to interaction of FS with the poloxamer chains.

Below their gelation temperature, poloxamer solutions exhibit Newtonian behaviour, whereby the shear stress is directly proportional to the shear rate.^[28] F1–F6 exhibited this behaviour at 20°C (Figure 5a) as they have gelation temperatures greater than 20°C. Viscosity increased with increase in poloxamer 237 concentration as shown by the steeper plots. This was likely due to increased hydrophobic interactions between poloxamer 407 and poloxamer 237 molecules. Previously, the gelation temperature poloxamer solutions exhibit pseudoplastic (shear thinning) behaviour, where viscosity decreases exponentially as a function of increasing shear stress and shear rate.^[29] These temperature dependent rheological changes ensure that the FS-poloxamer solutions remain as a free-flowing solutions making them easy for injection at room temperature, but once it is injected into the sclera (at physiological temperature), a more vicious gel phase is formed and remains at the site of injection, providing sustained drug release.

Figure 5.

Effect of 0.5% w/w fluorescein sodium loading on rheological behaviour of poloxamer mixtures containing 12% w/w poloxamer 407 and variable concentrations of poloxamer 237 (i.e. (F1) 15% w/w, (F2) 16% w/w, (F3) 17% w/w, (F4) 18% w/w, (F5) 19% w/w and (F6) 20% w/w of poloxamer 237) at (a) 20°C and (b) 37°C. n = 3.

It is essential to minimise the pain experienced by the patient upon injection, especially for invasive procedures such as intravitreal or periocular injections.^[32] Therefore, it is a common practice for eye drops containing local anaesthetics (often tetracaine or lidocaine in combination with adrenaline) to be instilled onto the surface of the eye before intravitreal injection to numb the injection site. The pain is due to two mechanical processes: the initial piercing and penetration of the scleral tissue, and then the injection of the solution into the vitreous humour that also raises intraocular pressure.

To overcome the pain and tissue trauma associated with ocular injections or surgical implantation, we have proposed a minimally invasive HMN device that can selectively localise the thermosensitive poloxamer polymers within the scleral tissue and result in forming intrascleral implants.

Before performing minimally invasive intrascleral injections of F1–F6 poloxamer formulations the maximum force and work (i.e. syringeability forces) required for injecting the formulations through the HMN device were investigated. The HMN gauge (i.e. 26, 29 and 30 G) and volumes (i.e. 30, 50 and 100 μl) tested in this study are commonly used for intravitreal injections. Initially, maximum forces required to overcome the frictional forces of the syringe to expel air from the HMN device into water was determined at 20°C. As shown in Tables 3–5, the maximum force required to expel air increased with volume of air and decreasing needle diameter (26–30 G). Similar results were observed for F1–F6 formulations when injected from 26, 29 and 30 G HMN devices, as shown in Table 3–5. Increasing formulation viscosity also increased the maximum force requirement. In parallel to the maximum force, the work required to inject FS-loaded formulations also increased with the volume of formulation and decreasing the needle diameter (Tables 6–8). The maximum force and work required to inject air were significantly higher than those for any formulation; this is due to the frictional forces associated with the syringes in injecting different volumes.

Table 3.

Maximum forces required to expel different volumes of air (i.e. to overcome the frictional forces) and formulations from hollow microneedle devices fabricated from 26 G needle into water at 20°C, n = 5

Formulation	Volume of formulation (μl)
	30	50	100
	Mean maximum force (±standard deviation) (in N)
Expel air	1.394 ± 0.048	2.125 ± 0.054	2.448 ± 0.033
F1	0.158 ± 0.026	0.325 ± 0.102	0.445 ± 0.060
F2	0.248 ± 0.028	0.449 ± 0.056	0.626 ± 0.059
F3	0.359 ± 0.050	0.600 ± 0.059	0.841 ± 0.068
F4	0.475 ± 0.034	0.823 ± 0.039	1.035 ± 0.078
F5	0.571 ± 0.044	1.099 ± 0.076	1.272 ± 0.047
F6	0.681 ± 0.024	1.329 ± 0.031	1.480 ± 0.078

Table 5.

Maximum forces required to expel different volumes of air (i.e. to overcome the frictional forces) and formulations from hollow microneedle devices fabricated from 30 G needles into water at 20°C, n = 5

Formulation	Volume of formulation (μl)
	30	50	100
	Mean maximum force (±standard deviation) (in N)
Expel air	2.000 ± 0.026	3.067 ± 0.062	3.601 ± 0.071
F1	0.401 ± 0.083	0.567 ± 0.047	1.011 ± 0.124
F2	0.529 ± 0.057	0.735 ± 0.064	1.252 ± 0.097
F3	0.694 ± 0.033	0.915 ± 0.056	1.436 ± 0.047
F4	0.910 ± 0.045	1.195 ± 0.091	1.654 ± 0.107
F5	1.131 ± 0.049	1.477 ± 0.044	1.836 ± 0.086
F6	1.301 ± 0.025	1.824 ± 0.012	2.021 ± 0.091

Table 6.

Work (i.e. syringeability) required to overcome the frictional forces (i.e. expel of air) and for expelling different volumes of formulations from hollow microneedle devices fabricated from 26 G needle into water at 20°C, n = 5

Formulation	Volume of poloxamer formulation (μl)
	30	50	100
	Mean syringeability (±standard deviation) (Ns)
Expel air	4.042 ± 0.048	7.257 ± 0.054	11.371 ± 0.083
F1	0.173 ± 0.047	0.268 ± 0.094	0.378 ± 0.095
F2	0.322 ± 0.054	0.493 ± 0.078	0.634 ± 0.061
F3	0.877 ± 0.034	0.974 ± 0.050	1.226 ± 0.055
F4	1.367 ± 0.042	1.476 ± 0.064	1.782 ± 0.066
F5	1.580 ± 0.033	1.817 ± 0.114	2.477 ± 0.093
F6	1.920 ± 0.067	2.342 ± 0.070	3.398 ± 0.024

Table 8.

Work (i.e. syringeability) required to overcome the frictional forces (i.e. expel of air) and for expelling different volumes of formulations from hollow microneedle devices fabricated from 30 G needles into water at 20°C, n = 5

Formulation	Volume of poloxamer formulation (μl)
	30	50	100
	Mean syringeability (±standard deviation) (Ns)
Expel air	6.009 ± 0.064	11.854 ± 0.098	15.696 ± 0.055
F1	0.496 ± 0.147	0.788 ± 0.084	1.714 ± 0.075
F2	1.054 ± 0.191	1.392 ± 0.097	2.236 ± 0.063
F3	1.684 ± 0.154	1.984 ± 0.066	2.798 ± 0.077
F4	2.250 ± 0.070	2.814 ± 0.103	3.767 ± 0.063
F5	3.263 ± 0.084	3.722 ± 0.084	4.835 ± 0.059
F6	4.519 ± 0.129	4.907 ± 0.110	6.000 ± 0.007

Table 4.

Maximum forces required to expel different volumes of air (i.e. to overcome the frictional forces) and formulations from hollow microneedle devices fabricated from 29 G needle into water at 20°C, n = 5

Formulation	Volume of formulation (μl)
	30	50	100
	Mean maximum force (±standard deviation) (in N)
Expel air	1.805 ± 0.029	2.963 ± 0.107	3.493 ± 0.044
F1	0.266 ± 0.049	0.394 ± 0.085	0.492 ± 0.076
F2	0.425 ± 0.053	0.567 ± 0.075	0.715 ± 0.036
F3	0.570 ± 0.027	0.764 ± 0.140	0.941 ± 0.141
F4	0.749 ± 0.047	0.993 ± 0.067	1.213 ± 0.080
F5	0.887 ± 0.043	1.048 ± 0.070	1.345 ± 0.060
F6	1.083 ± 0.055	1.215 ± 0.056	1.399 ± 0.061

Table 7.

Work (i.e. syringeability) required to overcome the frictional forces (i.e. expel of air) and for expelling different volumes of formulations from hollow microneedle devices fabricated from 29 G needle into water at 20°C, n = 5

Formulation	Volume of poloxamer formulation (μl)
	30	50	100
	Mean syringeability (±standard deviation) (Ns)
Expel air	5.640 ± 0.042	11.469 ± 0.082	15.207 ± 0.025
F1	0.443 ± 0.042	0.660 ± 0.081	1.223 ± 0.049
F2	0.816 ± 0.041	1.086 ± 0.056	1.772 ± 0.138
F3	1.394 ± 0.081	1.611 ± 0.088	2.404 ± 0.053
F4	1.883 ± 0.034	2.180 ± 0.052	3.216 ± 0.081
F5	2.679 ± 0.057	2.918 ± 0.066	4.051 ± 0.110
F6	3.360 ± 0.114	3.613 ± 0.047	4.741 ± 0.104

OCT, a non-invasive optical imaging technique, is often highlighted as the optical analogue to ultrasound. OCT maps the variation of reflected light rather than sound from a biological sample as a function of depth.^[33] Because of extensive light scattering of skin tissue, the typical penetration depth of optical techniques is low. However, OCT is capable of penetrating to depths of approximately 2.0 mm in comparison with confocal microscopy, which can only reach depths of approximately 250 μm,^[34] therefore, is the only optical method for cross-sectional imaging of the sclera and other ocular tissue layers in vivo and ex vivo. Thus, it has been used extensively for ophthalmological medical imaging of the retina and anterior segment of the eye, and it has also been reported in the use of calculating distances and permeation rates of solutes throughout the cornea and scleral tissues.^[35]

OCT showed an increase in scleral thickness around the circumference of the eye. A mean thickness of 422 μm was noted proximal to the lens, and near the orbit, the thickness was 531 μm and reached its maximal thickness of 675 μm proximal to the optic nerve. The penetration force required to pierce the sclera increased with depth of penetration and increasing needle's diameter, as represented in Table 9.

Table 9.

Mean penetration forces required to pierce rabbit scleral tissue using hollow microneedle devices with different needle sizes, n = 5

Needle gauge	Scleral location
	Equator (400 μma)	Anterior (500 μma)	Posterior (600 μma)
	Mean penetration force (±standard deviation) (in N)
26G	1.159 ± 0.084	1.346 ± 0.065	1.589 ± 0.101
29G	0.703 ± 0.034	0.886 ± 0.063	1.101 ± 0.098
30G	0.549 ± 0.077	0.710 ± 0.087	0.905 ± 0.058

Penetration depth.

In this study, for the first time, we have shown the ability of HMN device to precisely localise intrascleral implant-forming sustain release gel system and the scleral-pore closure. Measurements were made using image analysis software that confirmed the penetration of HMNs into the sclera to the required depth, creating a temporary void space in the sclera. When 50 μl of F6 formulation was injected into this void space, a gel-based implant was formed, as shown in Figure 6. The formed gel depot provided a sustained release of FS (Figure 7). F6 was chosen as it was the most viscous and had the lowest gelation temperature in the desired range, hence ensuring implant formation and retention within the void space intrasclerally. The most interesting observation was the reformation/closure of the sclera tissue, which appeared to envelope the implant encouraging retention in the injection site. Recovery time was dependent upon the depth of needle penetration. For example, recovery appeared to be complete within 2, 2.5 and 3 h for injection at a depth of 400, 500 and 600 μm (Figure 6), respectively. Further studies are required to assess the effect of long-term injection to the mechanical properties and pore-closure ability of the sclera.

Figure 6.

Optical coherence tomography images showing 30 G hollow microneedle injection of 50 μl F6 formulation (coloured in red) injected into equatorial sclera to a depth of (A) 400 μm at (a) 0, (b) 1 and (c) 2 h; (B) 500 μm at (a) 0, (b) 1, (c) 2 and (d) 2.5 h; and (C) 600 μm at (a) 0, (b) 1, (c) 2, (d) 2.5 and (e) 3 h. The arrow indicates empty space in sclera created following hollow microneedle application and its subsequent closure over time.

Figure 7.

In vitro release of fluorescein sodium from F1, F4 and F6 formulations at 37°C following intrascleral injection into rabbit sclera at (a) 400 μm (equatorially), (b) 500 μm (anteriorly) and (c) 600 μm (posteriorly), using 30 G hollow microneedle device. Mean ± standard deviation, n = 3.

Permeation of FS from the intrasclerally implanted gels was conducted using a modified Franz-cell setup (Figure 4). Depth of implant delivery within the sclera significantly affected %FS release over 24 h (Figure 7). The release data is best interpreted as the remaining distance that FS must diffuse through the scleral to reach the release media when the average scleral thickness of a given position is used as the maximum vertical distance. For example, when a 50 μl of F6 formulation was injected into equatorial sclera to a depth of 400 μm, then FS has to travel a mean distance of 22 μm before it can reach the receptor compartment of the Franz-diffusion cells. Similarly, FS has to travel a mean distance of 31 and 75 μm of the sclera following injections to a depth of 500 μm (anterior sclera) and 600 μm (posterior sclera), respectively. After 24 h, %FS release ranged from 80.88% to 87.56% (Figure 7a), 70.33% to 75.01% (Figure 7b) and 52.86% to 60.63% (Figure 7c) for an injection at 400 μm (equatorially), 500 μm (anteriorly) and 600 μm (posteriorly), respectively. Additionally, there was a small but significant difference between the %FS release between F1, F4 and F6 after 24 h. Clearly, these results indicate that the rate and extent of drug release following intrascleral delivery is dependent upon the region of scleral injection, depth of implantation and composition of formulation.

Conclusions

In conclusion, a minimally invasive HMN device was fabricated, with varying needle heights that enable precise localisation of in situ forming FS-loaded thermoresponsive poloxamer formulations within the sclera. Sustained release of FS was observed, and the percentage of FS release was depended upon the depth of implant delivery and region of intrascleral injection. Also, for the first time, this study investigated the pore-closure property of scleral tissue by using OCT, which was dependent upon the depth of HMN penetration. Finally, this study demonstrated a promising method of administering implant-forming gels in a minimally invasive manner, which could be a potential alternative to invasive intravitreal injections. Furthermore, sustained drug delivery can be achieved by varying the depth of needle penetration, composition of implant-forming gels and region of intrascleral injections. Further studies are now needed to investigate the effect of multiple intrascleral injections of implants by using HMN device and study its effect on drug release and scleral recovery; in addition, long-term implant retention should also be investigated. We also believe that HMN causes lesser scleral damage when compared with intravitreal injections; however, this needs further investigations.

Declarations

Conflict of interest

The Author(s) declare(s) that they have no conflicts of interest to disclose. None of the authors has a financial or proprietary interest in any method or material mentioned.