A Novel Amphotericin B Hydrogel Composed of Poly(Vinyl Alcohol)/Borate Complex for Ophthalmic Formulation

Abstract

In this study, we developed a water-soluble complex-hydrogel viscosity-controlled formulation of amphotericin B (AmB). AmB is insoluble in water, but borax makes it soluble by forming a complex with AmB. Borax also forms complexes with poly(vinyl alcohol) (PVA) to produce viscous hydrogels. Furthermore, boric acid interacts with mucin expressed in corneal epithelial cells. Accordingly, by utilizing these properties of borax simultaneously, we prepared a water-soluble AmB complex-hydrogel with poly(vinyl alcohol)/borate (PVA-B-AmB), which is suitable for eye drops. PVA-B-AmB was easily prepared by simply mixing aqueous AmB solution dissolved in borax, PVA solution, and water. The ¹¹B-NMR results suggested that PVA-B-AmB existed by bonding PVA and AmB via boronic acid. PVA-B-AmB (gel ratio = 0.55) has a viscosity of 18.3 ± 0.5 mPa·s and is suitable for ophthalmic formulations. This formulation exhibited sustained release of AmB of approximately 45% at 24 h. It was also shown that this formulation interacts with mucin. These results suggest that PVA-B-AmB can be used as a water-soluble AmB preparation suitable for ophthalmic use.

Introduction

Amphotericin B (AmB) is the gold standard for the treatment of fungal infections owing to its broad-spectrum antifungal and rapid fungicidal activity.¹⁾ AmB is insoluble in water and most organic solvents, posing challenges in its formulation.²⁾ Fungizone^® (AmB deoxycholate, dAmB) and lipid formulations have been developed for the treatment of systemic fungal infections. However, the use of dAmB and lipid formulations is limited due to side effects and cost, respectively.³⁾ These formulations are also used off-label for the treatment of superficial mycoses, such as fungal keratitis,⁴⁾ but have difficulties in terms of viscosity and mucosal adherence.⁵⁾ Therefore, an ophthalmic formulation of AmB with favorable properties for the treatment of superficial mycoses is required.

Poly(vinyl alcohol) (PVA) is a simple linear hydrophilic polymer. PVA is an economical and highly biocompatible material, and it has been used as a pharmaceutical additive and polymer-drug carrier.⁶⁾ In addition, PVA is rich in 1,3-diol linkages, which readily complex with dissociated borax ions in water to form viscous hydrogels.⁷⁾ It has also been reported that dissociated borax ions solubilize AmB in water by forming a complex with AmB.⁸⁾ In addition, boronic acid is known to form a complex with mucin, which is known to be expressed in the corneal epithelium, and is expected to be applied as an ophthalmic formulation for prolonging the corneal residence time.⁹⁾ However, to the best of our knowledge, there are no reports on formulations that simultaneously utilize these borax properties.

Therefore, we developed a new AmB formulation composed of PVA and borax (PVA-B-AmB) and the properties of PVA-B-AmB were investigated in vitro to determine its suitability for ophthalmic use.

Results and Discussion

Figure 1 shows ¹¹B-NMR spectra. Borax dissociates into boronic acid [B(OH)₃] and borate anion [B(OH)₄]⁻ in water. Boronic acid and borate anions are in ionization equilibrium, and their equilibrium is sufficiently rapid that they appear as a single NMR peak. This peak is determined between 1.6–19.3 ppm, depending on the ratio of boronic acid to borate anion present; 1 : 1 and 2 : 1 six-membered ring complexes of 1,3-diol and borate anions are reported to have a peak at 1.6 ppm and 1.2 ppm (Supplementary Figs. 1A and 1B), respectively.¹⁰⁾ The borax solution (4 mM, pH 10.5) exhibited a peak at 2.8 ppm. Poly(vinyl alcohol)/borax hydrogel (PVA-B) NMR spectrum showed a shift of the “free-boron”-derived peak to lower magnetic fields and an increased intensity of a new peak at 1.2–1.6 ppm derived from “complexed-boron” in a PVA concentration-dependent manner, respectively. Furthermore, depending on the PVA concentration, the 1.6 ppm-peak intensity decreased and the 1.2-ppm peak intensity increased (Fig. 1A). These results are consistent with those of previous studies,^10,11) suggesting that the number of 2 : 1 boron complexes (Supplementary Fig. 1B) increased with increasing PVA concentration.

Fig. 1. ¹¹B-NMR Spectra of (A) PVA-B, (B) B-AmB, and (C) PVA-B-AmB at Various Concentrations

Ethyl ether-boron trifluoride was used as the external reference 0.0. Data are shown from −10 to 20 ppm. B-AmB, amphotericin B dissolved in borax solution; PVA-B, poly(vinyl alcohol)/borax hydrogel; PVA-B-AmB, amphotericin B complex-hydrogel with poly(vinyl alcohol)/borax.

B-AmB NMR spectrum showed an AmB concentration-dependent shift of the peak to lower magnetic fields (from 2.8 to 5.0 ppm, Figs. 1A, 1B), whereas no new peaks were observed (Fig. 1B). The interaction of 1,2-diol with borate anions forms 1 : 1 and 2 : 1 five-membered rings complexes with peaks at 6 and 10 ppm (Supplementary Figs. 1C, 1D), respectively.¹¹⁾ In the measurement conditions, the concentration of AmB is much lower than that of borate anion, so these peaks of B-AmB NMR spectrum represent peaks of “free-boron” and five-membered ring complexes composed of AmB and borate anions (Fig. 1B).

PVA-B-AmB results showed a concentration-dependent increase in the intensity of the six-membered ring “complexed-boron” peaks (1.2 and 1.6 ppm, Fig. 1C), similar to PVA-B results (Fig. 1A). In addition, the peak observed in B-AmB (5.0 ppm, Fig. 1B) shifted to a low magnetic field depending on the concentration of PVA (5.8 and 12.0 ppm, Fig. 1C). It has been reported that the chemical shift of the spirocyclic complex consisting of five-membered and six-membered rings (Supplementary Fig. 1E) occurs in a magnetic field lower than 3 ppm.¹¹⁾ Therefore, PVA-B-AmB may be a spirocyclic complex consisting of a five-membered ring of B-AmB and a six-membered ring of PVA-B.

In summary, the NMR results suggest that PVA-B-AmB may be formed by crosslinking of borate anions with PVA or with AmB and PVA. In addition, UV-Visible spectrum results indicate that AmB in PVA-B-AmB exists in monomeric form (Supplementary Fig. 2 and Supplementary Table 1), which may have low cytotoxicity.¹²⁾ In fact, erythrocyte hemolysis experiments as a toxicity screen for AmB showed that PVA-B-AmB was less toxic compared with AmB and dAmB (Supplementary Fig. 3).

The optimal viscosity of ophthalmic solutions is recommended to be in the range of 15–50 mPa·s.¹³⁾ Contemporarily, viscosity should not exceed 30 mPa·s to avoid patient discomfort.¹⁴⁾ Figure 2 shows the viscosity of PVA-B-AmB and PVA-B at each gel ratio. The viscosity of PVA-B at gel ratios of 0.5, 0.55, and 0.6 were 10.1 ± 0.6, 18.0 ± 0.5, and 46.3 ± 1.2 mPa·s, respectively. The viscosity of PVA-B-AmB at gel ratios of 0.5, 0.55, and 0.6 were 10.0 ± 0.2, 18.3 ± 0.5, and 47.4 ± 0.9 mPa·s, respectively, indicating that the incorporation of AmB at a concentration of 0.1 mg/mL did not affect the viscosity. Based on the above, PVA-B-AmB with a gel ratio of approximately 0.55 was considered suitable for ophthalmic formulations.

Fig. 2. Viscosity Measurements of PVA-B and PVA-B-AmB at Gel Ratios of 0.5–0.6 (n = 3)

PVA-B, poly(vinyl alcohol)/borax hydrogel; PVA-B-AmB, amphotericin B complex-hydrogel with poly(vinyl alcohol)/borax.

The interaction of PVA-B-AmB with mucin tethered to the corneal and conjunctival epithelial cell membranes was pseudo-evaluated using mucin-coated nitrocellulose membranes (Fig. 3). The amount of AmB bound to membrane pretreated with PVA-B before exposure to PVA-B-AmB was significantly lower than that without pretreatment. Furthermore, exposure of the mucin and PVA-B-AmB mixture to mucin-coated nitrocellulose membranes resulted in negligible amounts of membrane bound AmB. These results suggest that some amount of PVA-B-AmB interact with mucin mimicking membrane-tethered mucin via PVA-B. Formulations that interact with mucin are thought to prolong corneal residence time.^15,16) Considering its viscosity and mucoadhesive properties, PVA-B-AmB is considered a potential ophthalmic formulation with prolonged corneal residence.

Fig. 3. Adhesion of PVA-B-AmB (Gel Ratio = 0.55) to Mucin-Coated Nitrocellulose Membrane (n = 3)

PVA-B, poly(vinyl alcohol)/borax hydrogel; PVA-B-AmB, amphotericin B complex-hydrogel with poly(vinyl alcohol)/borax. Difference between all groups is significant (p < 0.05).

To investigate the potential of PVA-B-AmB as an AmB drug delivery system, in vitro release experiments were performed (Fig. 4). Since mucin turnover is 15–20 h,¹⁷⁾ the amount of AmB released 24 h after the start of the experiment was compared. No significant difference was found between the PVA-B-AmB gel ratio = 0.5 and 0.55, but significant differences were found among the other groups. Water-soluble PVA-B-AmB, which did not exhibit suspension as shown in Supplementary Fig. 4, showed gel ratio-dependent sustained drug release. In addition, the initial release was lower for a higher gel ratio. Since boric acid has a pK_a of 9.27,¹⁸⁾ most of it exists as boronic acid [B(OH)₃] in phosphate-buffered saline (PBS) (pH 7.4). This makes it difficult to crosslink PVA, which may lead to degelation. It is also considered that AmB is released by degelation of PVA-B-AmB. Considering that the viscosity of PVA-B-AmB increases with increasing gel ratio, PVA-B-AmB with a high gel ratio is difficult to miscible with PBS and to decrosslink. A polymer network built up by dynamic covalent crosslinks may affect the release of AmB from PVA-B-AmB. And the release rate and amount of AmB can be adjusted according to its gel ratio. Therefore, PVA-B-AmB formulation that can interact with mucin and exhibits a sustained release of AmB may reduce the frequency of administration. AmB eye drops have been studied using concentrations of 30 µg/mL to 3 mg/mL.¹⁹⁾ Since this study investigated AmB concentrations from 25 µg/mL to 2 mg/mL, which is almost within the above range, the findings in this study are considered to be useful basic data for the development of PVA-B-AmB ophthalmic formulations.

Fig. 4. Release Curves of AmB from PVA-B-AmB at 37 °C

PVA-B, poly(vinyl alcohol)/borax hydrogel; PVA-B-AmB, amphotericin B complex-hydrogel with poly(vinyl alcohol)/borax. At 24 h, there are significant differences (p < 0.05) except for the groups with PVA-B-AmB gel ratio = 0.5 and 0.55.

Conclusion

In this study, we developed a water-soluble AmB formulation in which PVA and AmB are considered to be crosslinked via borate. Moreover, the formulation exhibits mucoadhesion and can be formulated to control both viscosity and drug release. Overall, this study revealed the in vitro properties of a complex-hydrogel system and its potential for future in vivo studies as ophthalmic formulations.

Experimental

Materials

PVA (MW: 80 kDa, degree of hydrolysis: 98.0–99.0%) was kindly supplied by the Japan Vam & Poval Co. (Osaka, Japan). AmB, dAmB, and mucin from porcine stomach type II were purchased from Cayman Chemical Co. (MI, U.S.A.), Bristol-Myers Squibb Co. (Tokyo, Japan), and Sigma-Aldrich Japan Co. (Tokyo, Japan), respectively. Other reagents and solvents were obtained from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan).

Preparation of PVA-B-AmB

PVA was dissolved in distilled water to prepare 4% (w/v) PVA solution. AmB solution (B-AmB) was obtained by dissolving AmB in 20 mM borax solution (pH 10.5). PVA-B-AmB was obtained by mixing 4% (w/v) PVA solution with B-AmB at a ratio of 4 : 1 (v/v). Similarly, PVA-B was obtained by mixing 4% (w/v) PVA solution with 20 mM borax solution (pH 10.5) at a ratio of 4 : 1 (v/v). This complex hydrogel was defined as having a gel ratio of 1. Samples with low gel ratios were prepared by appropriate dilution of B-AmB with distilled water and then mixing it with 4% (w/v) PVA solution. The preparations are shown in Supplementary Table 2.

¹¹B-NMR Analysis

Samples of borax solution, PVA-B, B-AmB, and PVA-B-AmB were prepared at various concentrations and measured by ¹¹B-NMR using JNM-ECZ400S (128 MHz; Jeol, Tokyo, Japan).

Viscosity Measurement

The viscosities of PVA-B-AmB (100 µg/mL AmB-equivalent) and PVA-B were measured using SV-10 Vibro Viscometer (A&D Company, Tokyo, Japan) at 25 ± 1 °C.

In Vitro Mucoadhesive Property

The interaction between mucin and PVA-B-AmB was examined according to a previously reported method, with slight modifications.²⁰⁾ A schematic protocol is described in Supplementary Fig. 5. Nitrocellulose membranes (Merck, Darmstadt, Germany), each of 1 cm², were treated with 0.05 M KH₂PO₄–NaOH buffer (pH = 8.0) containing 0.15% Tween80 or 4% mucin and 0.15% Tween80. The membranes were then washed with PBS and, if necessary, incubated with PVA-B (gel ratio = 0.55) for 20 min. The membranes were then incubated with PVA-B-AmB containing 0.5% mucin or PVA-B-AmB (500 µg/mL AmB-equivalent, gel ratio = 0.55) for 20 min. After washing again, each membrane was dissolved in 500 µL dimethyl sulfoxide, and AmB was quantified by measuring the absorbance at 416 nm (UV-1800; Shimadzu, Kyoto, Japan).

Drug Release Assay

In vitro drug release assay of PVA-B-AmB was performed by dialysis with slight modifications.^21,22) PVA-B-AmB (4 mL) containing 100 µg of AmB at a gel ratio of 0 (B-AmB), 0.5, 0.55, or 1 was prepared. PBS at pH 7.4 containing 2% sodium deoxycholate was used as the release medium. A dialysis bag (Spectra/Por 7; 25 kD; Repligen, MA, U.S.A.) containing each sample was immersed in a beaker containing 96 mL of the medium. The vessels were covered and stirred at 120 rpm in a water bath at 37 °C. At appropriate time intervals, 1 mL of the medium was collected and replaced with an equal volume of fresh medium. AmB was quantified in the collected medium by measuring the absorbance at 416 nm.

Statistical Analysis

Student’s t-test was used to compare differences between two experimental groups. Differences between multiple groups were analyzed using ANOVA followed by Tukey’s multiple comparison test. The mean ± standard deviation for three independent experiments is shown, and statistical significance was set at p < 0.05. Statistical tests were performed using the R software (version 4.1.0; Vienna, Austria).

Acknowledgments

The authors thank Japan Vam & Poval Co. for providing PVA. We appreciate the valuable advice provided by Dr. Tetsuya Nakamura. We thank Chiaki Fujita and Nanami Toba for their technical assistance.

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

This article contains supplementary materials.

References

• 1) Faustino C., Pinheiro L., Pharmaceutics, 12, 29 (2020).
• 2) Torrado J. J., Espada R., Ballesteros M. P., Torrado-Santiago S., “Amphotericin B formulations and drug targeting,” John Wiley and Sons Inc., Hoboken, 2008.
• 3) Morand K., Bartoletti A. C., Bochot A., Barratt G., Brandely M. L., Chast F., Int. J. Pharm., 344, 150–153 (2007).
• 4) Chhonker Y. S., Prasad Y. D., Chandasana H., Vishvkarma A., Mitra K., Shukla P. K., Bhatta R. S., Int. J. Biol. Macromol., 72, 1451–1458 (2015).
• 5) Göttel B., Lucas H., Syrowatka F., Knolle W., Kuntsche J., Heinzelmann J., Viestenz A., Mäder K., Front. Bioeng. Biotechnol., 8, 600384 (2020).
• 6) Rowe R. C., Sheskey P. J., Quinn M. E., “Handbook of Pharmaceutical Excipients,” 6th ed., Pharmaceutical Press, London, U.K., 2009.
• 7) Rezvan G., Pircheraghi G., Bagheri R., J. Appl. Polym. Sci., 135, 46734 (2018).
• 8) Strauss G., Kral F., Biopolymers, 21, 459–470 (1982).
• 9) Liu S., Jones L., Gu F. X., Macromol. Biosci., 12, 1622–1626 (2012).
• 10) Sinton S. W., Macromolecules, 20, 2430–2441 (1987).
• 11) Henderson W. G., How M. J., Kennedy G. R., Mooney E. F., Carbohydr. Res., 28, 1–12 (1973).
• 12) Cuddihy G., Wasan E. K., Di Y., Wasan K. M., Pharmaceutics, 11, 99 (2019).
• 13) Račić A., Čalija B., Milić J., Milašinović N., Krajišnik D., J. Drug Deliv. Sci. Technol., 49, 50–57 (2019).
• 14) La Gatta A., Corsuto L., Salzillo R., D’Agostino A., De Rosa M., Bracco A., Schiraldi C., J. Ocul. Pharmacol. Ther., 34, 677–684 (2018).
• 15) Bongiovì F., Fiorica C., Palumbo F. S., Di Prima G., Giammona G., Pitarresi G., Mol. Pharm., 15, 5031–5045 (2018).
• 16) Ludwig A., Adv. Drug Deliv. Rev., 57, 1595–1639 (2005).
• 17) Franca J. R., Fuscaldi L. L., Ribeiro T. G., Foureaux G., Cesar A. L. A., Castilho R. O., Cronemberger S., Ferreira A. J., Fernandes S. O. A., Cardoso V. N., Faraco A. A. G., J. Biomed. Mater. Res. B Appl. Biomater., 108, 2227–2237 (2020).
• 18) “CRC Handbook of Chemistryand Physics, Internet Version 2005,” ed. by Lide D. R., CRC Press, Boca Raton, FL, 2005.
• 19) Kaushik S., Ram J., Brar S. G., Jain A. K., Chakraborti A., Gupta A., Cornea, 20, 715–719 (2001).
• 20) Banshoya K., Fujita C., Hokimoto Y., Ohnishi M., Inoue A., Tanaka T., Kaneo Y., Int. J. Pharm., 610, 121061 (2021).
• 21) Yang Z. L., Li X. R., Yang K. W., Liu Y., J. Biomed. Mater. Res. A, 85A, 539–546 (2008).
• 22) Jain J. P., Kumar N., Eur. J. Pharm. Sci., 40, 456–465 (2010).