Supercritical Drug Impregnation onto Intraocular Lenses Abir BOULEDJOUIDJA a, Yasmine MASMOUDI a, *, Baoguang JIANG b,c,wei HE b,c, Elisabeth BADENS a a Aix Marseille université, CNRS, Centrale Marseille, M2P2 UMR 7340, 13451, Marseille, France b Shenyang Bio Medical Device Co., Ltd., Shenyang 110163, Liaoning Province, China c Shenyang He Eye Hospital, Shenyang 110034, Liaoning Province, China * Yasmine.masmoudi@univ-amu.fr Abstract Supercritical impregnation is an attractive clean alternative to conventional soaking into liquid process, using generally organic solvents. Impregnation process can be used for the development of controlled delivery systems that can be applied in pharmaceutical and medical fields. This work focuses on the preparation of controlled drug delivery systems for ophthalmic applications through supercritical impregnation. More particularly, commercial intraocular lenses (IOLs) used in cataract surgery are loaded with ciprofloxacin (an antibiotic) and dexamethasone 21-phosphate disodium salt (an anti-inflammatory drug) in order to prevent short- and long-term postoperative complications. Foldable hydrophilic and rigid hydrophobic acrylate IOLS, made respectively from derivative of poly (2-hydroxyethyl methacrylate) (P-HEMA) and Poly (methyl methacrylate) PMMA were studied. Supercritical impregnations were carried out in a batch mode and impregnation yields were determined through drug release kinetic studies in a solution simulating the aqueous humor. The influence of operating conditions on impregnation was studied by performing preliminary impregnation experiments followed by experimental designs. Transparent IOLs presenting an effective impregnation were obtained. The highest impregnation yields for DXP and CIP in PMMA IOLs were 18.3 and 2.8 µg drug /mg IOL respectively and the highest impregnation yields for DXP and CIP in P-HEMA IOLs were 14.5 and 4.1 µg drug /mg IOL respectively. Those results indicate higher affinity of DXP for PMMA and P-HEMA IOLs than CIP, which was confirmed by partition coefficient values. Despite the low solubility of each drug in the fluid phase, homogeneous and in-depth impregnations were successfully obtained. A prolonged drug delivery during 40 days was obtained for most impregnation experiments (for both kinds of IOLs). Optical characterizations of IOLs were carried out by evaluating the diopter power as well as the imaging quality through the determination of the Modulating Transfer Function at a specified spatial frequency according to European Standard (ISO 11979-2:2014). Results show that optical properties were conserved after supercritical impregnation. I. Introduction Among other diseases, cataract is the most common cause of blindness and severe visual impairment worldwide. The number of patients with cataract is continuously increasing [1]. It is conventionally treated through a surgery consisting in replacing the opacified natural crystalline lens with a synthetic intraocular lens (IOL). It is generally safe but the risk of postoperative endophthalmitis has to be considered and an IOL implantation is always a concern even with topical drug coverage [1]. A relevant solution to overcome those drawbacks could be the use of controlled drug delivery systems (DDS) placed inside the eye. These DDS can allow a slow release of drug over time in the potential infection area [2]. If the DDS is the impregnated IOL, this solution does not require an additional act on the surgeon. As an alternative to conventional impregnation techniques, supercritical fluid impregnation, especially using supercritical carbon dioxide (scco 2 ) can reduce or even eliminate the use of organic solvents [3]. Furthermore, when applied to polymers, scco 2 can act as a swelling and/or plasticizing agent promoting therefore the impregnation process. 1
The aim of the present work was to study the supercritical impregnation of commercially available intraocular lenses (IOLs), used for cataract surgery with an anti-inflammatory drug (dexamethasone 21-phosphate disodium salt, namely DXP) and an antibiotic (ciprofloxacin, designed CIP). Two polymeric IOLs were particularly tested: rigid IOLs (PMMA) and foldable IOLs (hydrated in their original form) (P-HEMA). A number of parameters such as pressure, impregnation duration and the presence/quantity of co-solvent were investigated to achieve optimized drug loading, homogeneous distribution of drugs in IOLs and a controlled drug release. II. Materials Supercritical impregnation was performed on two types of commercially available IOLs supplied by the Fred Hollows Intraocular Lens (Nepal): Rigid hydrophobic IOLs made from derivative of Poly (methyl methacrylate) (PMMA): IOLs with a diopter of +21.0 D were used, Foldable IOLs made from derivative of Poly (2-hydroxyethyl methacrylate) (P-HEMA): IOLs are initially conditioned in a soaking physiological solution. IOLs with three different diopters were used (+5.0D, +21.0D and 32.0D). Ciprofloxacin (CIP, C 17 H 18 FN 3 O 3, 516 g.mol -1 ) and dexamethasone 21-phosphate disodium salt (DXP, C 22 H 28 FNa 2 O 8 P, 331 g.mol -1 ) are commonly used ophthalmic drugs in postoperative cataract treatment. Both drugs were supplied by Sigma-Aldrich (France). The solvents employed in this work were carbon dioxide (99.7 % purity) from Air Liquide (France) and ethanol ( 99.8 % purity) supplied by Groupe MERIDIS (France). III. Methods III. 1. Experimental design and response surface methodology In RSM, an empirical mathematical model (Eq. 1) is postulated and a suitable experimental design is performed to estimate required coefficients. This model, once validated can be used to predict the response in the whole experimental domain with good precision [4]. k k y = b 0 + b i x i + b ii x i 2 + b ij x i x j i=1 j=1 i<j Eq. 1 In a first step, a series of preliminary supercritical impregnation was carried out in order to delimit the experimental operating domain by varying different parameters. Based on the obtained results, two different experimental designs were elaborated for each type of IOLs, considering two factors in a central composite design with 9 individual design points in a spherical domain. III. 2. Supercritical impregnation set-up A detailed description of the batch impregnation process is presented elsewhere [5]. 2
IV. Characterizations IV. 1. Drug release kinetics studies In-vitro drug release kinetics studies were conducted by immersing impregnated IOLs in 5 ml of simulated aqueous humor (ph of 7.2) under stirring in a closed vessel at 310 K. An aliquot of 0.4 ml was collected every day for 60 days and drug (CIP or DXP) release was quantified through spectrophotometer analyses (Jenway 6715 UV/Vis spectrophotometer). Aliquots were then returned to the release vessel to maintain the initial volume. IV. 2. Impregnation yields The impregnated amounts were determined through release studies and defined as cumulative mass of the released drug after reaching a constant value. The impregnation yield was calculated according to the following equation (Eq. 2) where, m imp is the mass of impregnated API (corresponding to m during release) and m 0 IOL is the initial mass of dry IOL. Y imp = m imp m 0 IOL Eq. 2 IV. 3. Partition coefficients The affinity between the drug and the polymer can be determined through the partition coefficient (K), corresponding to the mass ratio of the drug in the impregnation matrix to that in scco 2 phase. It is calculated with equation 3, where y drug-polymer and y drug-co2 are the mass fractions of the drug in the impregnation matrix and in the fluid phase respectively. K = y drug polymer y drug CO2 Eq. 3 IV. 4. Optical properties The diopter power and the modulating transfer function (MTF) at a specified spatial frequency of IOLs were characterized according to the European Standard methods (ISO 11979-2:2014). V. Results and discussions In this work, we are interested in the impregnation of drugs into PMMA and P-HEMA IOLs. For each kind of IOL, the results were discussed first for DXP and then for CIP. Solubility measurements of both drugs (DXP and CIP) in scco 2 (without using a co-solvent) were carried out at 308 K [6]. Experimental conditions and drug solubilities are summarized in the Table 1. Solubilities of both DXP and CIP in scco 2 are low, with higher values obtained for CIP. An increase in pressure (from 8 to 20 MPa) enhances drug solubility in scco 2 (1.22 10-7 to 2.15 10-7 for DXP and 1.83 10-7 to 4.51 10-7 for CIP in molar fraction). To our knowledge experimental data of DXP and CIP in scco 2 is presented for the first time in this work. Table 1. Solubilities of DXP and CIP (y drug in molar fraction) in scco 2 at 308 K [6]. P (MPa) CO 2 density (kg.m -3 ) y drug DXP CIP 8 419.1 1.22 10-7 1.83 10-7 14 801.4-3.11 10-7 20 865.7 2.15 10-7 4.51 10-7 3
V. 1. PMMA IOLs V. 1. 1. Dexamethasone 21-phosphate disodium salt impregnation a. Preliminary impregnation In a first part of this study, supercritical impregnations were performed with pure scco 2 or using a co-solvent (ethanol, 5 %mol) at 8 and 20 MPa on IOLs with a diopter of +21.0 D (Table 2). Temperature was kept constant at 308 K, impregnation duration at 2 hours and depressurization rate at 0.2 MPa.min -1. Experiment repeatability was verified by reproducing impregnations three or four times. Table 2. Influence of pressure and the use of a co-solvent (ethanol 5 %mol) on supercritical impregnation of PMMA IOLs with DXP (at 308 K and for 2 hours). N m 0IOL (*) Y Pressure imp t release m DXP imp (**) (mg) (MPa) ( g) ( g drug /mg IOL) days Without co-solvent DXP_1 18.9 0.2 8 159 24 8.4 1.3 40 DXP_2 19.0 0.2 20 165 24 8.7 1.3 40 With co-solvent (5 %mol) DXP_3 20.0 0.2 8 240 36 12.0 1.8 44 DXP_4 20.1 0.2 20 99 15 4.9 0.7 40 (*) Initial mass of the dry IOL before impregnation; (**) Release duration In the absence of co-solvent, pressure increase from 8 to 20 MPa had no significant effect on impregnation yields. Addition of ethanol at low pressure (8 MPa) resulted in an increase in impregnated DXP amount. It is well known that a co-solvent such as ethanol promotes the solubility of polar drugs as well as CO 2 sorption in polymers and swelling/plasticizing effect. However, a lower impregnation yield was obtained at 20 MPa which can be explained by a drug partition becoming more favorable towards the fluid phase. Interestingly, impregnation yields were improved at low pressure in presence of co-solvent. Hence, all DXP loadings later in this work were performed at 8 MPa and in the presence of co-solvent. b. Experimental design According to these first results, a response surface methodology based on experimental designs was used to study the influence of the amount of the co-solvent (1 to 10 %mol) as well as the impregnation duration (30 to 240 min). The pressure was kept at 8 MPa, the temperature at 308 K and the depressurization rate at 0.2 MPa.min -1. Figure 1 shows impregnated amounts and impregnation yields evolution respectively predicted by the RSM model in terms of amount of co-solvent and impregnation duration defined as the two input variables of the experimental design. The amount of impregnated drug varies between 118 and 398 g corresponding to impregnation yields between 5.6 to 18.3 g drug /mg IOL, which suggests that the studied factors influence the surface response. According to the response surface presented in Figure 1, the predominant effect is the impregnation duration. Indeed, response variations are significant in the defined experimental domain, particularly for low quantities of co-solvents (lower than 5.5 %mol). The most 4
advantageous conditions of impregnation are short duration (30 min) with low amount of cosolvent (5.5 %mol). µg 320 m imp 210 100 Figure 1. Two-dimensional contour plot of impregnation yields and three-dimensional response surface of impregnated mass for supercritical impregnation of PMMA IOLs with DXP. V. 1. 2. Ciprofloxacin impregnation Supercritical impregnation of CIP in PMMA IOLs was carried out at various conditions, the influence of pressure (8 and 20 MPa) and use of a co-solvent (5 %mol ethanol) on drug impregnation was studied for diopter +21.0 D. Other parameters i.e. temperature (308 K), impregnation duration (2 hours), depressurization rates (0.2 MPa.min -1 ) were kept constant for all experiments. The obtained results were presented in the table 3. In the absence of co-solvent, increase in pressure from 8 to 20 MPa improved significantly the impregnation yields (from 0.8 to 2.4 g drug /mg IOL ). This could be attributed to the concurrent increase in CIP solubility in scco 2 (from 1.83 10-7 to 4.51 10-7 respectively), CO 2 sorption in the polymeric IOLs and their resulting swelling. Addition of ethanol as co-solvent in the procedure resulted in a further increase in drug impregnation yield (from 0.8 to 2.8 g drug /mg IOL at 8 MPa). However, in presence of ethanol, increase in pressure (from 8 to 20 MPa) did not affect the drug loading. This could be either due to the saturation of IOLs with drug or no further improvement in polymer/drug interactions at higher pressures. Table 3. Influence of pressure and the use of a co-solvent (ethanol 5 %mol) on supercritical impregnation of PMMA IOLs with CIP (at 308 K and for 2 hours). N m 0IOL Pressure m CIP imp Y imp t release (mg) (MPa) ( g) ( g drug /mg IOL) (days) Without co-solvent CIP_1 19.0 0.2 8 16 1 0.8 0.1 12 CIP_2 19.6 0.2 20 48 4 2.4 0.2 40 With co-solvent CIP_3 19.9 0.2 8 55 5 2.8 0.0 40 CIP_4 20.5 0.2 20 57 5 2.8 0.2 40 In-vitro kinetic studies of drug release from impregnated IOLs with each drug, were conducted for 60 days on all impregnated IOLs. Drug release for all the studied conditions x 2 x 1 : Amount of co-solvent x 2 : Impregnation duration t imp 5
exhibited a sustained release profile (about 40 days) without an initial burst release which suggests deep impregnation of the drug in IOLs. NMR analyses were carried out on impregnated IOLs in the presence of co-solvent without and with carrying out a CO 2 washing step phase (1 hour) before depressurization. Ethanol peaks (non shown in this article) disappear when the washing phase is carried out indicating an efficient removal of ethanol (residual ethanol lower than 0.01 wt %) which is suitable for the intended application. V. 2. Comparison between PMMA and P-HEMA IOLs impregnation Supercritical impregnation study of P-HEMA IOLs with both drugs (DXP and CIP) is described elsewhere [5]. In order to compare impregnation of both IOLs referring to partition coefficients, results obtained with both drugs in similar experimental conditions in the absence of co-solvent are presented in table 4. Table 4. Comparison of PMMA and PHEMA IOLs impregnation with CIP and DXP in absence of a co-solvent. Pressure Y imp ( g drug /mg K Y imp ( g drug /mg IOL) K (MPa) IOL) PMMA P-HEMA DXP 8 8.4 1.3 5.88 10 3 8.5 1.3 5.94 10 3 20 8.7 1.3 3.44 10 3 9.1 1.4 3.61 10 3 CIP 8 0.8 0.1 6.11 10 2 0.95 0.1 7.22 10 2 20 2.4 0.2 7.21 10 2 2.86 0.3 8.44 10 2 Despite low solubilities of DXP and CIP (slightly higher for CIP) in scco 2, partition coefficients were higher for DXP indicating a better affinity than CIP with both polymeric matrixes. Regarding DXP impregnation, the partition coefficient (K) decreases with pressure increase from 8 to 20 MPa for both types of IOLs indicating a partition of the drug becoming more favorable towards the fluid phase. This phenomenon can be explained by two factors. First, an increase in CO 2 density with pressure results in increased solvent power of CO 2, enhancing thus DXP solubility in the fluid phase (1.22 10-7 to 2.15 10-7 in molar fraction). At the same time, increasing CO 2 density promotes PMMA swelling favoring DXP dragging in the fluid phase during depressurisation. The partition coefficients of DXP were quite similar for both IOLs, which could be explained by quite similar affinities between DXP and both IOLs (PMMA and P-HEMA) in these conditions. Regarding CIP impregnation, in absence of co-solvent, the partition coefficient (K) increases with pressure increase (8 to 20 MPa) in similar way for both types of IOLs. These results could be explained by an increase in CIP solubility in the fluid phase (1.83 10-7 to 4.51 10-7 ) and a weak affinity between CIP and both kinds of IOLs. In absence of co-solvent, the partition coefficients of CIP were quite comparable for both types IOLs indicating by quite similar affinities between CIP and both types of IOLs. 6
V. 3. Optical properties In order to study the influence of supercritical impregnation on the optical properties of IOLs, optical characterization were performed. IOLs (P-HEMA and PMMA blends) commercialized by Shenyang Bio Medical Device were used for this part of study since they present well defined initial optical properties. Diopter power and modulating transfer function (MTF) at a specified spatial frequency of this type of IOLs were determined before and after supercritical treatment or impregnation with different diopters according to the European Standard methods and conditions (ISO 11979-2:2014). IOLs optical properties respect the European standard conditions if the MTF values are higher or equal to 0.43 in both directions (sagittal and meridional) and the tolerance limits on spherical diopter power are respectively of ( 0.4) and ( 0.5) for the IOLs with diopters between (15 and 25) and (25 and 30). The experimental conditions and corresponding optical properties of IOLs treated with scco 2 and impregnated with DXP and CIP are presented respectively in table 5 and 6. Optical properties were conserved either after supercritical treatment and/or supercritical impregnation. Indeed, the values of diopter and MTF respect the standard conditions which is suitable for the intended application. Table 5. Influence of the supercritical treatment on the optical properties of IOLs. Before treatment After treatment Pressure Ethanol (MPa) (%mol) Diopter MTF MTF MTF MTF Diopter (sagittal) (meridional) (sagittal) (meridional) 1 8 0 24.62 0.487 0.463 24.53 0.495 0.445 2 8 5 24.34 0.498 0.481 24.45 0.518 0.476 3 20 0 22.73 0.541 0.493 22.82 0.533 0.489 4 20 0 17.11 0.547 0.512 17.1 0.553 0.499 Table 6. Influence of the supercritical impregnation on the optical properties of IOLs. Before impregnation After impregnation Drug Pressure Ethanol MTF MTF MTF MTF Diopter Diopter (MPa) (%mol) (sagital) (meridional) (sagital) (meridional) 1 CIP 20 0 18.52 0.46 0.45 18.55 0.45 0.45 2 CIP 20 0 28.41 0.57 0.55 28.342 0.56 0.57 3 DXP 8 0 26.98 0.54 0.51 26.837 0.53 0.49 4 DXP 8 0 20.67 0.47 0.48 20.518 0.49 0.47 5 DXP 8 0 13.01 0.53 0.52 12.969 0.52 0.49 6 DXP 8 5 23.53 0.54 0.53 23.421 0.54 0.52 VI. Conclusion This work aimed to load rigid and foldable intraocular lenses (PMMA and P-HEMA) with ophthalmic drugs in order to combine cataract surgery and postoperative treatment in a single procedure. Two commonly used drugs to prevent cataract postoperative complications, ciprofloxacin and dexamethasone 21-phosphate disodium salt, were studied in this work. Supercritical impregnation was carried out in a batch mode and impregnation yields were determined through drug release studies for both types of IOLs. Despite the low solubility of drugs (DXP and CIP) in the fluid phase, homogeneous / in depth impregnated IOLs and presenting a prolonged drug delivery during 40 days were successfully obtained for most impregnation experiments (both kinds of IOLs). 7
Higher impregnation yields were obtained for both IOLs (PMMA and P-HEMA) with DXP than with CIP indicating better affinity of DXP with both IOLs than CIP and this result was confirmed by the values of partition coefficient. NMR analyses were performed on impregnated lenses (PMMA and P-HEMA) with both drugs in the presence of a co-solvent and did not show presence of residual ethanol. These results showed that impregnated IOLs are suitable for ophthalmic application. An important result of this work is that optical properties of IOLs were maintained after supercritical treatment and impregnation. Such results have never been reported until now. References [1] EPERON, S., RODRIGUEZ-ALLER, S., BALASKAS, K., GURNY, R., and GUEX- CROSIER, Y., International Journal of Pharmaceutics, vol. 443, 2013, p. 254. [2] SAURINA, J., DOMINGO, C., KAZARIAN, S. G., ANDANSON, J. M., GARCÍA- GONZÁLEZ, C. A., LÓPEZ-PERIAGO, A., FERNÁNDEZ, V., and ARGEMÍ, A., Journal of Supercritical Fluids, vol. 48, 2009, p. 56. [3] BADENS, E., Techniques de l ingénieur, 2012, vol. 33. [4] BAŞ, D and BOYAC İ. H., Journal of Food Engineering, vol. 78, 2007, p. 836. [5] BOULEDJOUIDJA, A., MASMOUDI, Y., SERGENT, M., TRIVEDI, V., MENIAI, A., and BADENS, E., International Journal of Pharmaceutics, vol. 500, 2016, p. 85. [6] CHEN, Y-P., Personal communication, National Taiwan University, 2015. 8