Photo-Responsive Hydrogel Microneedles With Interlocking Control for Easy Extraction in Sustained Ocular Drug Delivery

Age-related macular degeneration (AMD) and diabetic retinopathy (DR) are the leading causes of blindness in the United States [1,2]. Both AMD and DR are caused in part by over production of vascular endothelial growth factor (VEGF). In the presence of excessive VEGF, the capillaries begin to leak causing large molecules to form exudates to escape into the retina. This leakage causes edema in the surrounding tissues or abnormal growth of vessels. Clinical trials using VEGF inhibitors have shown superior results compared to laser photocoagulation for retinal and macular diseases [3,4]. Anti-VEGF therapies have been the standard treatment method for AMD and DR [5–8]. Delivery of anti-VEGF agents needs to be localized to reduce systemic adverse effect [9] and, therefore, is generally administered via intravitreal injections to bypass barrier properties of ocular tissues [10,11]. The intravitreal elimination half-life of anti-VEGF drugs, such as Ranibizumab and Bevacizumab, is around 4–6 days. Thus, in order to maintain sufficient VEGF binding activities, regular intravitreal injections are required. Most clinical trials use a treatment regime with regular monthly injections over the course of 2 years. However, intravitreal injections have been associated with complications including endophthalmitis, intra-ocular inflammation, retinal detachment, intra-ocular pressure elevation, and intra-ocular hemorrhage [12].

Microneedles (MNs) offers a platform for minimally invasive drug delivery and are being investigated as an alternative to intravitreal injections [13–15]. Hydrogels have diverse and great polymer swelling characteristics and thus have been used with great success to achieve sustained drug delivery due to their ability to slow the diffusion of drugs into target systems [16–18]. Thus, many researchers opted to use hydrogels as the fabrication material of their MNs [19–21]. It has been shown that the MN can be used for sustained drug release, with a zero-order release profile, for four to six weeks, making it a promising alternative to intravitreal injections [22]. However, one challenge for using MNs for ocular drug delivery is fixation of the MN, because of the sensitivity of the human eye and the limited area available for trans scleral drug delivery. In transdermal applications of MN, adhesive backing is often used, but this is not applicable in the eye. Chemical adhesives present inflammatory side effects to the eye [23]. Hydrogel adhesives use covalent bonds to surface biomolecules but only provide limited adhesion [24]. Sutures and staples offer firm adhesion but can lead to scarring and tissue damage.

A bio-inspired MN design with an interlocking or barbed feature has been proposed to address the issue of adhesion [25]. The interlocking feature, taking advantage of the swellable feature of hydrogels, can help the MN to self-adhere to application site without needing any adhesives. In a previous study, a fabrication process of MN with interlocking features was developed and demonstrated that the self-adhesive MN shows a significant increase in adhesion strength without increasing the required penetration force [15].

Yet, MNs with interlocking features present another challenge, i.e., the removal of the MN. While the bio-inspired interlocking features increase the adhesion strength, they are not designed to be easily removed without causing pain or tissue damages. A high extraction force may cause retinal tear or detachment as well as an increased intra-ocular pressure which can lead to damages to the optic nerve. In order to apply MNs with interlocking features to the eye for self-adhering, there needs a way to detach the MN array from the eye. The goal of this study is to investigate a strategy for easy removal of the self-adhered MNs. In this study, we propose to use a photoresponsive hydrogel in which its swelling and deswelling can be controlled through the stimulation of light, as shown in Fig. 1. Spiropyrans are a series of classic photochromic dyes which have been widely used in the development of stimuli-responsive materials [26,27]. Spiropyran functionalized N-isopropylacrylamide (NIPPAM) hydrogels have been developed as a photoresponsive material [28] and used in microfluidic applications as values and actuators [29,30]. When the hydrogel is subject to light, the hydrogel switches a hydrophilic ring-open form to a hydrophobic ring-closed form because of the spiropyran. This hydrophobic nature will cause deswelling or shrinking of the hydrogel. By employing this feature, we can reverse the MN adhesion by inducing deswelling for easy extraction.

In this study, we first developed a fabrication process for the spiropyran conjugated hydrogel MNs with three different compositions (0%, 10%, and 20% of spiropyran-conjugated poly(NIPPAM) gels (PNS) while the remaining is polyvinyl alcohol (PVA)). The swelling and deswelling kinetics were characterized. The penetration force and adhesion strengths with and without light stimulation were compared to evaluate the performance of the photoresponsive hydrogel MNs.

The MNs consist of two material systems, one with PVA which provides the mechanical strength of the MN and the other with the spiropyran conjugated PNS. Polyvinyl alcohol (PVA) hydrogel (M_w 146,000–186,000, 99% hydrolysis; Sigma Aldrich, St. Louis, MO) was used with a concentration of 16% PVA to water w/w. The PVA and water mixture was stirred at 95 °C until the PVA was fully dissolved. Then the mixture was vacuumed and degassed before mixing with the spiropyran conjugated PNS. The spiropyran was prepared using the reported procedure [28–30]. ¹H-NMR of spiropyran was measured (400 MHz, CDCl₃): δ 7.17 (t, J = 7.60 Hz, 1H), 7.06 (d, J = 7.2 Hz, 1H), 6.83 (t, J = 7.4 Hz, 1H),6.79 (d, J = 10.2 Hz, 1H), 6.68-6.58 (m, 3H), 6.51 (d, J = 7.7 Hz, 1H), 6.40 (d, J = 17.3 Hz, 1H), 6.12 (dd, J = 17.3, 10.5 Hz, 1H), 5.82 (d, J = 10.4 Hz, 1H), 5.69 (d, J = 10.2 Hz, 1H), 4.17 (t, J = 6.7 Hz, 2H), 3.89 (t, J = 6.4 Hz, 2H), 2.72 (s, 3H), 1.81-1.66 (m, 4H), 1.53-1.39 (m, 4H), 1.30 (s, 3H), 1.16 ppm (s, 3H). This NMR data resemble reported data. PNS was prepared by polymerization of spiropyran and N-isopropylacrylamide (NIPAAM) in the present of azobisisobutyronitrile, following the reported procedure [28–30]. ¹H-NMR (400 MHz, DMF-d₇): δ 7.50 (br, 35H), 7.14 (dd, J = 7.40 Hz, 3H), 7.05 (d, J = 10.4 Hz, 1H) 6.88-6.84 (br, 1H), 6.80 (t, J = 7.4 Hz, 1H), 6.75 (d, J = 8.1 Hz, 1H), 6.58 (dd, J = 13.4, 8.4 Hz, 2H), 5.83 (d, J = 9.6 Hz, 1H), 3.96 (s, 45H), 2.27-2.16 (m, 40H), 1.81-1.38 (m, 80H), 1.13 ppm (s, 244H). These resemble reported data.

The fabrication process follows a prior study by the authors [15] with the addition of the PNS to the mixture of the PVA hydrogel as shown in Fig. 2. The positive master molds were fabricated using a digital light processing-based three-dimensional printer (B9 Core 530; B9Createions, Rapid City, SD). Next, elastomer SYLGARD^® 184 (Dow Corning, Midland, MI) was cast into the master molds to create the fabrication molds (negative) as demonstrated by Bediz et al. [31] where a ratio of 1:10 curing agent to elastomer was used and the mixture was centrifuged at 2300 rpm for a duration of 5 min in a centrifuge (NuWind NU-C200R; NuAire, Plymouth, MN). Three hydrogel compositions (0%, 10%, and 20% of spiropyran-conjugated PNS) were used to fabricate MN arrays. The hydrogel mixtures were mixed by a magnetic stirrer. Finally, the mixture was cast over the fabrication mold, centrifuged, and subjected to seven freeze-thaw cycles. In each cycle, the mixture is subjected to freezing for 8 h at −20 °C and then thawing for 5 h at 25 °C [32,33].

Figure 3 shows the experimental setup used to conduct the force versus distance measurement tests. The setup consists of a linear motion system converted from a three-dimensional printer (TAZ 6, LulzBot, Fargo, ND), a load cell (Miniature Low Profile Tension Link Load Cells 0.75 to 1" Height, OMEGA, Norwalk, CT), a data acquisition system (Model No. USB-1608G, Measurement Computing, Norton, MA) and a light source (OmniCure S2000 Spot Curing System, Excelitas Technologies, Waltham, MA). The linear motion system was used to drive the MNs into the polycaprolactone-based sclera-mimicking phantom, which was fabricated following the procedures outlined in a previous study by the authors [15]. The phantom was fixed on an acrylic holder. The load cell was used to measure the penetration and adhesion forces. Three MN arrays of different compositions were used as follows: 100% PVA, 90% PVA- 10% PNS (w/w), and 80% PVA-20 PNS (w/w).

To measure the penetration force, the MN was lowered to right above the sclera-mimicking phantom and then was programed to advance 1.5 mm at a speed of 100 mm/min. An example of the recorded force is shown in Fig. 4. At first, the MN deflected the phantom until point A where the tip of MN cut through and penetrated into the phantom. At point B, the interlocking features penetrated into the phantom. The maximum recorded force at point C was used as the measure of the insertion force.

To measure the extraction force without light illumination, the inserted MNs stayed in for 10 min to allow swelling, and then they were extracted by moving the MN holder up a rate of 100 mm/min. An example of the recorded extraction force is also shown in Fig. 4. The force at point D was used as the extraction force for each measurement. In the cases of deswelling experiments, the MNs were allowed to swell for 10 min and then underwent light illumination using a light source, as shown in Fig. 3, was used to illuminate the MN for 14 min at 10 mW/cm² irradiation of 365 nm before the MN was extracted.

To measure the swelling in the vitreous humor mimicking gel and deswelling under light illumination, an inverted microscope was used to take images of an MN, and ImageJ was used to measure the change in area and the width of the interlocking feature. The MNs were inserted into the vitreous humor mimicking gel contained in an acrylic box and sealed with a 1-mm layer of the sclera-mimicking phantom.

One-way analysis of variation (ANOVA) was performed using EXCEL to examine the statistical significance of the differences of extraction and insertion forces among the 100% PVA, 90% PVA- 10% PNS, and 80% PVA-20 PNS MN arrays.

Figure 5 shows the average results of ten extraction tests each on hydrogel composition of 80% PVA/20% PNS (w/w) and 90% PVA/10% PNS (w/w) MNs with and without light illumination for deswelling. The extraction force is defined as the maximum measured force when the MNs inserted into the testing material are being pulled out from the sclera-mimicking phantom. The results show a significant difference between the extraction force with and without light illumination for the 80% PVA/20% PNS MNs. This study demonstrated a reduced extraction force can be achieved. However, it would still require an animal study to validate that the reduced extraction force leads to less tissue damage when the MN is being removed.

Figure 6 shows a comparison of the extraction force in two groups, one with light illumination and the other without light illumination. In each group, the results of three different MN compositions (100% PVA, 10% PNS/90% PVA, and 20% PNS/80% PVA) are shown. Without light illumination, the 20% PNS/80% PVA MNs show the highest extraction force, while with light illumination, the same MNs show the lowest extraction force, signaling easy extraction due to deswelling of the MNs. MNs of both 100% PVA and 10% PNS/90% PVA show insignificant change in extraction force with and without light illumination. A ultraviolet (UV) light source was used in this study, as the PNS is shown previously to be most responsive to this wavelength. To protect the eye while applying the illumination, a white light frequency can be used instead of a UV frequency to avoid potential damage to the retina. This volume reduction induced by the light illumination can be further optimized in the future with better mixtures of PNS and PVA

Figure 7 compares the insertion forces of ten averaged results for three MN compositions. The insertion force is defined as the maximum recorded force when an MNs was inserted into the test material while no light illumination was introduced. The results show that the insertion force has a slight increase as the composition of PNS increases. This could be due to the weakening of the mechanical properties of the hydrogel with the addition of PNS which is not as strong as PVA. With lower mechanical properties, the MNs do not cut through the tissue easily and are more susceptible to buckling or bending while penetrating the tissue. While it is expected that with a higher fraction of spiropyran infused hydrogel, a better deswelling behavior can be achieved, but this benefit comes with the tradeoff of a minor increase in insertion force, yet, the penetration force is still within acceptable limits, below 5 N.

Figure 8 shows the swelling and deswelling of an MN (20% PNS/80% PVA) over time. In the first 10 min, the MN was submerged into a vitreous-mimicking fluid and swelled. Light illumination was applied to the MN at the 10 min mark for 14 min and the deswollen geometry was shown. The width of the interlocking feature and the total area of the MN was used to quantify the swelling/deswelling kinetics, as shown in Fig. 9. At 10 min, the width of the interlocking feature and the total area of the MN increased by 70% and 50%, respectively. Upon light illumination, about 20% reduction of MN in both the width and total area was observed. No significant reduction was observed by continuing illuminating the MNs. This deswelling of MN contributes to the reduction of adhesion force, as shown in Figs. 2 and 3. As shown in Fig. 5, a higher concentration of PNS leads to a more significant reduction in extraction force. However, increasing the PNS content also leads to an increase in insertion force, as shown in Fig. 7. An optimal composition of the MNs with a balanced performance of both insertion and deswelling will require further investigation. On the other hand, a potential strategy to mitigate this is to selectively increase the fraction of PNS locally at the interlocking feature. Therefore, one strategy to further improve the performance of the MN is to selectively increase the PNS content at the interlocking feature. This will require a multistep manufacturing process but has been shown to be feasible [34].

In this study, we demonstrated a photoresponsive MN that can self-adhere to the tissue upon swelling and can deswell for easy extraction after illuminated with light. A critical composition of the spiropyran conjugated hydrogel is required to enable the photoresponsive feature. With 20% spiropyran conjugated hydrogel, the width of the interlocking feature reduces by 20% and subsequently, a nearly 20% reduction in the required force to remove the MN.

In this experiment, the barb-like interlocks was chosen to cause moderate interlocking with smooth angles of entry and exit, yet if further interlocking is needed the angels can be made more aggressive and that will further emphasize the need for a volume reduction of the needles to allow extraction. Although the difference in the extraction force reduction between PVA and PVA-PNA hydrogels is not big, yet this pilot study shows the potential of volume reduction which has been shown to reduce extraction forces. With further optimization of the hydrogel composition and light illumination the need for the PVA-PNA hydrogel will be more significant.

This new concept opens the door to many possibilities of sustained drug delivery. The device was able to lock itself with no need for external adhesives, sutures or staples. This will allow for longer sustained drug delivery and easy removal, without the need of surgical removal, when drug delivery is completed. This allows for use in different locations in the human body where chemical adhesives use will not be possible due to bodily fluids.

This research was funded by U.S. Department of Defense - Congressionally Directed Medical Research Programs (W81XWH-18-1-0137). The authors would also like to acknowledge Drs. Nina Woodford, John DenHerder, and Gay Lynn Clyde of the Washington State University Office of the Campus Veterinarian for their help in this research and providing animal tissues.

• Congressionally Directed Medical Research Programs (Grant No. W81XWH-18-1-0137; Funder ID: 10.13039/100000090).