Photodegradation of the H1 Antihistaminic Topical Drugs Emedastine, Epinastine, and Ketotifen and ROS Tests for Estimations of Their Potent Phototoxicity

pharmaceutics

Article Photodegradation of the H1 Antihistaminic Topical Drugs Emedastine, Epinastine, and Ketotifen and ROS Tests for Estimations of Their Potent Phototoxicity

Anna Gumieniczek 1,* , Anna Berecka-Rycerz 1, Urszula Hubicka 2 , Paweł Zmudzki˙ 3 , Karolina Lejwoda 1 and Paweł Kozyra 1

1 Department of Medicinal Chemistry, Medical University of Lublin, Jaczewskiego 4, 20-090 Lublin, Poland; [email protected] (A.B.-R.); [email protected] (K.L.); [email protected] (P.K.) 2 Department of Inorganic and Analytical Chemistry, Jagiellonian University, Collegium Medicum, Medyczna 9, 30-688 Cracow, Poland; [email protected] 3 Department of Medicinal Chemistry, Jagiellonian University, Collegium Medicum, Medyczna 9, 30-688 Cracow, Poland; [email protected] * Correspondence: [email protected]

Received: 27 May 2020; Accepted: 15 June 2020; Published: 17 June 2020

Abstract: In this study, important H1 antihistaminic drugs, i.e., emedastine (EME), epinastine (EPI), and ketotifen (KET), were irradiated with UV/Vis light (300–800 nm) in solutions of different pH values. Next, they were analyzed by new high performance liquid chromatography (HPLC) methods, in order to estimate the percentage of degradation and respective kinetics. Subsequently, ultra-performance liquid chromatography tandem-mass spectrometry (UPLC-MS/MS) was used to identify their photodegradation products and to propose degradation pathways. In addition, the peroxidation of linoleic acid and generation of singlet oxygen (SO) and superoxide anion (SA) were examined, together with the molar extinction coefficient (MEC) evaluation, to estimate their phototoxic risk. The photodegradation of all EME, EPI, and KET followed pseudo first-order kinetics. At pH values of 7.0 and 10.0, EPI was shown to be rather stable. However, its photostability was lower at pH 3.0. EME was shown to be photolabile in the whole range of pH values. In turn, KET was shown to be moderately labile at pH 3.0 and 7.0. However, it degraded completely in the buffer of pH 10.0. As a result, several photodegradation products were separated and identified using the UPLC-MS/MS method. Finally, our ROS assays showed a potent phototoxic risk in the following drug order: EPI < EME < KET. All of these results may be helpful for manufacturing, storing, and applying these substantial drugs, especially in their ocular formulations.

Keywords: H1 antihistaminics; topical formulations; photodegradation kinetics; photodegradation pathways; phototoxicity; reactive oxygen species

1. Introduction The sensitivity of an active pharmaceutical substance (API) to light may vary with its chemical structure and reactivity, as well as the proposed dosage forms. For many APIs and formulations, photoreactivity leads to chemical degradation or to changes in their physical appearance [1,2]. Many classes of drugs also have the potential to provoke phototoxic, photoallergic, and photogenotoxic eﬀects in tissues exposed to light [1,3]. Emedastine (EME), epinastine (EPI), and ketotifene (KET) belong to H1 histaminic receptor antagonists, exhibiting varied activity and several speciﬁc properties. EME is known as a competitive

Pharmaceutics 2020, 12, 560; doi:10.3390/pharmaceutics12060560 www.mdpi.com/journal/pharmaceutics Pharmaceutics 2020, 12, x FOR PEER REVIEW 2 of 24

Pharmaceutics 2020, 12, 560 2 of 22 Emedastine (EME), epinastine (EPI), and ketotifene (KET) belong to H1 histaminic receptor antagonists, exhibiting varied activity and several specific properties. EME is known as a blockercompetitive of the blocker H1 receptor, of the asH well1 receptor, as an inhibitor as well as of calcium an inhibitor ion inﬂux of calcium across ion the influx basophil across plasma the membranebasophil plasma or intracellular membrane calcium or intracellular ion release calcium [4]. EPIion release has properties [4]. EPI has which properties inhibit thewhich release inhibit of mediators,the release such of mediators, as histamine such and as leukotrienes. histamine and These leukotriene includes. both These antihistamine include both properties antihistamine and chemicalproperties mediator and chemical stabilizer mediator properties stabilizer [4]. KET properties blocks the H[41].receptor KET blocks as a non-competitive the H1 receptor blocker, as a stabilizesnon-competitive mast cells, blocker, inhibits stabilizes the platelet mast cells, activating inhibits factor, the andplatelet acts activating as an eosinophil factor, and inhibitor acts as [4 ,an5]. Theeosinophil chemical inhibitor structures [4,5 of]. The the drugschemical of intereststructures are of shown the drugs in Figure of interest1. are shown in Figure 1.

N CH3 CH3 N N N N S N

O N NH2 O

CH 3 EME EPI KET Figure 1. Chemical structures of emedastine (EME), epinastine (EPI), and ketotifen (KET). Figure 1. Chemical structures of emedastine (EME), epinastine (EPI), and ketotifen (KET).

Blocking of the histamine H1 receptors serves as a primary therapeutic goal for topical anti-allergic medications.Blocking Therefore,of the histamine EME, EPI, H1 receptors and KET, serves which as are a used primary in eye therapeutic drops, external goal for solutions, topical andanti- ointments,allergic medications. can be mainly Therefore used to, EME,treatmany EPI, and allergic KET, problems which are[5]. usedOn the in other eye drops, hand, topically external appliedsolutions APIs, and mayointments, degrade canupon be mainly extensive used to light treat exposure, many allergic lowering problems their [ therapeutic5]. On the other action hand, or generatingtopically applied toxic products.APIs may Indegrade this context, upon extensive the examination light exposure, of their lowering photodegradation their therapeutic constitutes action a matteror generating of great toxic interest products. in the modernIn this context, pharmacy the examination field [6]. It is of worth their noting photodegradation that photodegradation constitutes isa notmat onlyter of related great interest to changes in the in modern the structure pharmacy of API, field but [ also6]. It to isthe worth occurrence noting that of free photodegradation radical processes, is energynot only transfer, related orto evenchanges luminescence, in the structure which of mayAPI, leadbut also to unexpected to the occurrence results of [7 ,free8]. radical processes, energyA reviewtransfer of, or older even APIs luminescence, in the area which of photostability may lead to raises unexpected some concerns, results [7,8 as]. the data collected so farA could review be of incomplete older APIs in in the the lightarea ofof newphotostability guidelines. raises Moreover, some concerns, the development as the data of collected modern analyticalso far could methods be incomplete has increased in the the light possibility of new ofguidelines. the identification Moreover, of new the degradationdevelopment products of modern [7]. Whenanalytical it comes methods to studies has increas on theed the stability possibility of EME, of the EPI, identification and KET, theof new literary degradation resources products are rather [7]. limited.When it Acomes literature to studies survey revealson the thestability presentation of EME, of EPI a densitometry, and KET, thin-layerthe literary chromatography resources are rather (TLC) stability-indicatinglimited. A literature method survey in reveals which the the presentation acidic and oxidative of a densitometry degradation thin of- EMElayer was chromatography described [9]. However,(TLC) stability there-indicating is no data method concerning in which the sensitivity the acidic to and UV/ Visoxidative light exposure. degradation In the of EME literature was concerningdescribed [9 EPI,]. However, there are severalthere is HPLCno data methods concerning for its the determination sensitivity to in UV/Vis bulk drug light substances, exposure. tablets, In the andliterature ocular concerning drops [10– 14 EPI]., Inthere these are previous several studies, HPLC EPImethods was degraded for its determination in acidic, alkaline, in bulk neutral, drug andsubstance oxidatives, tablets conditions, and ocular as a partdrops of [10 validation.–14]. In these In twoprevious studies studies, [12,13 EPI], EPI was was degraded also exposed in acidic, to UValkaline, light, neutral but the, and breadth oxidative of the conditions experiments as a waspart of rather validation. limited In and two no studies quantitative [12,13], resultsEPI was were also presented.exposed to A UV few light, HPLC but stability-indicating the breadth of the methods experiments have been was rather presented limited for theand determination no quantitative of KETresults in di werefferent presented. formulations A few [15 – HPLC18]. However, stability degradation-indicating of methods the drug have under been UV /Vis presented irradiation for hasthe notdetermination been examined of KET thus in far different and photodegradation formulations [15 kinetics–18]. However, has not been degradat described.ion of the drug under UV/VisTherefore, irradiation we carried has not out been a detailed examined investigation thus far on and the photostability photodegradation of EME, kinetics EPI, and has KET not under been variousdescribed. pH conditions. Next, HPLC and UPLC-MS/MS experiments were performed for quantitative determinationsTherefore, andwe kineticcarried measurements,out a detailed asinvestigation well as for theon separationthe photostability and identification of EME, EPI of, respective and KET photodegradationunder various pH products. conditions. Since Next, the literature HPLC and does UPLC not contain-MS/MS any experiments similar publications, were performed the studies for presentedquantitative here determinations are the first reports and in kinetic this regard. measurements, In addition, as the well photochemical as for the properties separation of EME, and EPI,identification and KET were of respective examined photodegradation in different ways in products. terms of Since their phototoxicity. the literature Firstly, does not we contain propose any an approachsimilar publications involving UV, the/Vis studies measurements presented from here290 are tothe 700 first and reports determination in this regard. of the molarIn addition, extinction the coephotochemicalfficient (MEC) properties values. of Subsequently, EME, EPI, and the KET colorimetric were examined monitoring in different of linoleic ways acid in terms peroxidation, of their phototoxicity. Firstly, we propose an approach involving UV/Vis measurements from 290 to 700 and

Pharmaceutics 2020, 12, 560 3 of 22 as well as singlet oxygen (SO) and superoxide anion (SA) generation in the presence of irradiated drugs, are reported.

2. Materials and Methods

2.1. Materials The pharmaceutical grade standards of emedastine difumarate (EME), epinastine hydrochloride (EPI), ketotifen hydrogen fumarate (KET), chlorprotixene and todralazine hydrochlorides (internal standards for HPLC methods, i.s.), quinine hydrochloride and benzocaine (positive and negative controls for ROS determinations), imidazole, linoleic acid, p-nitrosodimethylaniline (RNO), nitroblue tetrazolium chloride (NBT), thiobarbituric acid (TBA), butylhydroxytoluene (BHT), 1,10,3,30-tetraethoxyoxypropane, 1-butanol, dimethyl sulphoxide (DMSO), and methanol from Merck (Darmstad, Germany); glacial acetic acid, sodium acetate, hydrochloric acid, sodium chloride, sodium tetraborate, phosphoric acid, sodium hydrogen phosphate, sodium dihydrogen phosphate, kalium dihydrogen phosphate, and kalium hydroxide from POCh (Gliwice, Poland); and formic acid, ammonium formate, acetonitrile, and water for LC/MS from J.T. Baker (Center Valley, PA, USA), were used. The buffer solutions (acetate buffers of pH 3.0 and 4.8, phosphate buffers of pH 2.5 and 7.0, and borate buffer of pH 10.0) were prepared as described in European Pharmacopoeia [19]. The buffers of pH 3.0, 7.0, and 10.0 that were used as degradation media had the same ionic strength of 1 M, which was attained with 4 M sodium chloride. Sodium phosphate buffer of pH 7.4 (20 mM) that was used for the ROS assays was prepared according to respective literature [20].

2.2. Preliminary Spectrophotometric Measurements EME, EPI, and KET were dissolved in methanol (1 mg/mL) and then diluted with 20 mM sodium phosphate buffer of pH 7.4 to a final concentration of 10 µg/mL. The absorption spectra in the range of 200–700 nm were recorded with a UV/Vis double beam spectrophotometer (CE 6600 from Cecil Instruments Ltd., Milton, UK) and a quartz cell with a 1 cm path length. The molar extinction coefficients (MEC) were determined from the absorbance values for the peaks at the maximum wavelengths.

2.3. Solar Simulator The Suntest CPS Plus chamber from Atlas (Linsengericht, Germany) was used as a solar simulator irradiating UV and Vis light in the range of 300–800 nm, with the indicator setting value of 250 W/m2 for 1 h. The chamber was equipped with an appropriate temperature control unit and the temperature was adjusted at 25 ◦C during all experiments.

2.4. Photodegradation of EME, EPI, and KET in Solutions Volumes of 1 mL from the stock solutions of EME, EPI, or KET (1 mg/mL) were dispensed to quartz glass-stoppered dishes and diluted with 1 mL of respective buﬀer solution (pH 3.0, 7.0, and 10.0). The samples were exposed to UV/Vis light with energies equal to 18,902, 37,804, 56,706, 75,608, and 94,510 kJ/m2. These doses of light were attained during 7, 14, 21, 28, and 35 h of irradiation, respectively. The starting energy of 18,902 kJ/m2 was equivalent to 1,200,000 lux h and 200 W/m2, · which is recommended by the ICH Q1B guidelines [21] as an initial dose of light that conﬁrms drug stability, while the next doses were 2–5 times higher. After irradiation, the samples were diluted with methanol to cover the linearity range of our HPLC methods and analyzed quantitatively as described below. The assays were repeated three times for each sample (n = 3), and the concentrations of non-degraded drugs for each level of degradation (from 7 to 35 h of irradiation) were calculated from respective calibration equations and the starting concentration of EME, EPI, or KET. Pharmaceutics 2020, 12, 560 4 of 22

Kinetics of Photodegradation The concentrations of non-degraded drugs remaining after each time of irradiation or logarithms of these concentrations were plotted against the time of degradation to obtain the equation y = ax + b and the determination coeﬃcient r, and in consequence, to determine the reaction order. Finally, further kinetic parameters, i.e., the degradation rate constant (k), degradation time of the 10% substance (t0.1), and degradation time of the 50% substance (t0.5), were calculated.

2.5. HPLC Method

2.5.1. Chromatography Analysis was performed with a model 515 pump, a Rheodyne 20 µL injector, and a model UV 2487 DAD, controlled by Empower 3 software, all from Waters UK Sales (Elstree, England). For EME and EPI, chromatography was carried out on a LiChrospher®100CN column (125 4.0 mm, 5 µm) × from Merck. The mobile phase for EME was a mixture of methanol and phosphate buffer of pH 2.5 (40:60, v/v) with the flow rate of 2 mL/min. The UV detection was set at 280 nm. The mobile phase for EPI was the mixture of acetonitrile-methanol-acetate buffer of pH 4.8 (45:35:20, v/v/v) with the flow rate of 1.5 mL/min. The UV detection was set at 240 nm. For KET, chromatography was carried out on a LiChrospher®100RP-18 column (125 4.0 mm, 5 µm) from Merck. The mobile phase was a mixture of × acetonitrile-methanol-acetate buffer of pH 4.8 (30:40:30, v/v/v) containing 0.25% formic acid with the flow rate of 2.0 mL/min, while the UV detection was set at 296 nm.

2.5.2. Validation of the Methods The selectivity of the methods was examined by determination of the stressed drugs in the presence of their degradation products. For calibration, working solutions of EME, EPI, or KET were prepared by dispensing 0.1–1.0 mL volumes from the stock solutions (1 mg/mL) to 10 mL volumetric flasks in order to reach the concentration range of 10–100 µg/mL. Internal standards (i.s.) were added as follows: 0.6 mL volumes of the stock solution of chlorprotixene (1 mg/mL) for EME, 0.2 mL volumes of the stock solution of todralazine (1 mg/mL) for EPI, and 0.5 mL volumes of chlorprotixene solution (10 mg/mL) for KET. After adjusting with methanol to the mark, solutions were injected into the column in five repetitions. The ratios of peak areas (the drug and respective i.s.) were plotted against the corresponding concentration of the drug to construct the calibration equations. The limit of detection (LOD) and the limit of quantification (LOQ) were determined from the SD of the intercept and the slope of the respective regression lines. In order to verify the accuracy and repeatability, replicate (n = 3) injections of working solutions of EME and EPI at low (15 µg/mL), medium (50 µg/mL), and high (90 µg/mL) concentrations were conducted. In the case of KET, the concentrations of 30, 50, and 70 µg/mL were injected into the column. The accuracy was calculated as the percentage of the analyte recovered by the respective assay, while repeatability was calculated as the relative standard deviation for the repeated between-day estimations.

2.6. UPLC-MS/MS Analysis The UPLC-MS/MS system consisted of a Waters Acquity UPLC system coupled with a Waters TQD mass spectrometer (an electrospray ionization mode ESI-tandem quadrupole (Waters Corporation, Milford, MA, USA)). Chromatographic separations were carried out using the Acquity UPLC BEH C18 column (2.1 100 mm, 1.7 µm), equipped with the Acquity UPLC BEH C18 VanGuard pre-column × (2.1 5 mm, 1.7 µm) from Waters. The column was maintained at 40 C and eluted under gradient × ◦ conditions using from 95% to 0% of eluent A over 10 min, at a flow rate of 0.3 mL/min. Eluent A was water/formic acid (0.1%, v/v) and eluent B was acetonitrile/formic acid (0.1%, v/v). Chromatograms were recorded using the Waters eλ PDA detector. Spectra were analyzed in the 200–700 nm range with a 1.2 nm resolution and sampling rate of 20 points/s. The MS detection settings of the Waters TQD mass spectrometer were as follows: source temperature, 150 ◦C; desolvation temperature, 350 ◦C; Pharmaceutics 2020, 12, 560 5 of 22 desolvation gas flow rate, 600 L/h; cone gas flow, 100 L/h; capillary potential, 3.00 kV; and cone potential, 30 V. Nitrogen was used for both nebulizing and drying gas. The data were obtained in a scan mode ranging from 50 to 1000 m/z at 0.5 s intervals; eight scans were summed up to get the final spectrum. Collision-activated dissociation (CAD) analyses were carried out with the energy of 50 eV. Consequently, the ion spectra were obtained by scanning from 50 to 500 m/z. The data acquisition software employed was MassLynx V 4.1 from Waters.

2.7. ROS Assays For all ROS assays, EME, EPI, and KET were dissolved in a mixture of DMSO 20 mM sodium − phosphate buffer of pH 7.4 (1:1, v/v) to obtain the 10 mM stock solutions. The indicator setting value of 250 W/m2 for 1 h was used for our CPS Plus chamber, achieving an intensity of UVA from 6.5 to 7.9 J/cm2. Our samples were irradiated for 15, 30, and 45 min and for 1, 3, and 7 h, which corresponded to energies equal to 675, 1350, 2025, 2700, 8100, and 18,902 kJ/m2. Lipid peroxidation was measured using a TBA reactive substances (TBARs) test [22]. Briefly, linoleic acid was dissolved in 20 mM buffer of pH 7.4 (1 mM) and irradiated without or in the presence of EME, EPI, and KET (200 µM). Then, the irradiated samples (500 µL) were mixed with equal volumes of TBA (0.67% in 20 mM sodium phosphate buffer of pH 7.4) and 10 µL of BHT (1% in glacial acetic acid), and heated at 95 ◦C for 30 min. Finally, the samples were extracted with 1 mL of 1-butanol and centrifuged at 4000 rpm for 10 min, and the absorbance of the extracts was measured at 532 nm. The measurements were repeated three times (n = 3). A standard curve for TBARs was obtained with malondialdehyde (MDA) that had been generated by the hydrolysis of 1,10,3,30-tetraethoxypropane in acidic medium at a high temperature of 95 ◦C[22]. The calibration procedure with MDA as a standard was repeated five times (n = 5). Singlet oxygen (SO) determination was based on bleaching p-nitrosodimethylaniline (RNO) using imidazole as a selective acceptor of SO [20,22,23]. The samples containing EME, EPI, and KET (200 µM) were mixed with RNO (50 µM) and imidazole (50 µM) in the quartz glass-stoppered dishes. The absorbance of each sample was measured at 440 nm. Then, the samples were exposed to UV/Vis light as described above, and the absorbance at 440 nm was measured again. In the superoxide anion test (SA), the reduction of nitroblue tetrazolium (NBT) was monitored as an increase in the absorbance at 560 nm [20,22,23]. The samples containing EME, EPI, and KET (200 µM) and NBT (50 µM) were mixed thoroughly, measured at 560 nm, exposed to UV/Vis light, and measured again. All measurements were repeated three times for each sample (n = 3) and according to the results, the photoreactivity for EME, EPI, and KET was judged [20].

3. Results and Discussion

3.1. UV Measurements As the primary event of a phototoxic reaction, the substance is excited by the absorption of photon energy, and the photoexcited chemical can then react with other substances through energy transfer and/or radical reactions, possibly leading to phototoxicity [3,22]. Therefore, a UV/Vis spectral analysis is recommended by the oﬃcial guidelines as the ﬁrst step of photosafety testing in the pharmaceutical industry [24]. A UV/Vis spectral analysis to clarify the potent photoreactivity of EME, EPI, and KET was carried out over the range of 200–700 nm. According to the spectral patterns of each drug, intense UV absorption was observed within the tested range (Figure2). What is more, EME and KET showed at least one peak near or above 290 nm, partly overlapping with the sunlight spectrum, with MEC values greater than 10,000 L/(mol cm). The UV absorption spectrum of EME was characterized by · strong bands around 210, 255, and 290 nm. For the lowest energy band at 286 nm, the MEC value was calculated as 10,962 L/(mol cm). EPI was shown to absorb appreciably over the range of 200–240 nm, · while above 240 nm, the absorbance decreased gradually to 300 nm. The UV absorption spectrum of KET was characterized by two sharp peaks around 210 and 300 nm. For the lowest energy band at Pharmaceutics 2020, 12, 560 6 of 22

296 nm, the MEC value was calculated as 15,617 L/(mol cm). On the basis of these results, EME and · KET werePharmaceutics recognized 2020, 12, as x FOR probably PEER REVIEW being excitative by sunlight and photoreactive, as well as potentially6 of 24 phototoxic [23].

Pharmaceutics 2020, 12, x FOR PEER REVIEW 7 of 24

Figure 2. UV spectra of the examined drugs (concentration of 0.01 mg/mL in methanol). Figure 2. UV spectra of the examined drugs (concentration of 0.01 mg/mL in methanol).

3.2. Validation of the HPLC Method Validation of our HPLC methods was performed with respect to the selectivity, linearity, LOD and LOQ, precision, and accuracy, according to ICH guidelines [25]. The respective validation parameters are shown in Table 1.

Table 1. Validation of HPLC methods for the determination of emedastine (EME), epinastine (EPI), and ketotifen (KET) (n = 5).

Values Parameter EME EPI KET Linearity range (µg/mL) 10–100 10–100 10–100 Slope 0.01942 0.03020 0.0061 SD of the slope 0.00129 0.00028 0.00002 Intercept 0.06199 0.04752 0.0018 SD of the intercept 0.00913 0.00475 0.00012 r2 0.9991 0.9962 0.9997 SD of the r2 0.00011 0.00042 0.00018 LOD (µg/mL) 1.41 0.47 0.06 LOQ (µg/mL) 4.70 1.57 0.20 Accuracy (% Recovery) 97.95 104.65 102.69 SD of the recovery 1.73 1.14 2.10 Inter day precision (% RSD) 2.06 1.03 0.21

Pharmaceutics 2020, 12, 560 7 of 22

Table 1. Validation of HPLC methods for the determination of emedastine (EME), epinastine (EPI), and ketotifen (KET) (n = 5).

3.3. Percentage of Degradation of EME, EPI, and KET The photodegradation of EME, EPI, and KET was estimated with our HPLC methods, after exposing the drugs to UV/Vis light with energies equal to 18,902, 37,804, 56,706, 75,608, and 94,510 kJ/m2 (exposure for 7, 14, 21, 28, and 35 h, respectively). As a result, EME was observed to be photolabile in a wide pH range, with a degradation value of 32.38–41.52% after the maximal irradiation. Pharmaceutics 2020, 12, x FOR PEER REVIEW 8 of 24 The chromatogram shows a decrease of the peak corresponding to the non-degraded EME (tR 0.88 min). However, the appearance of any new peaks was not detected at 280 nm (Figure3A). As far as EPI is the peaks of degradation products in respective chromatograms (Figure 3B). When KET was concerned, the photodegradation was lower, achieving 8.10–19.10% with the pH decreasing from 10 to irradiated with the maximal dose of light, it degraded to 14.64% at pH 3.0 and to 19.28% at pH 7.0. 3 after the highest irradiation (94,510 kJ/m2), without a corresponding rise in the peaks of degradation However, almost 100% degradation of KET was observed at pH 10.0 after a much lower dose of products in respective chromatograms (Figure3B). When KET was irradiated with the maximal dose energy of 37,804 kJ/m2. The respective chromatogram shows the disappearance of the peak of light, it degraded to 14.64% at pH 3.0 and to 19.28% at pH 7.0. However, almost 100% degradation corresponding to unmodified KET (tR 0.88 min), as well as the appearance of several new peaks of of KET was observed at pH 10.0 after a much lower dose of energy of 37,804 kJ/m2. The respective degradation products (Figure 3C). chromatogram shows the disappearance of the peak corresponding to unmodiﬁed KET (tR 0.88 min), as well as the appearance of several new peaks of degradation products (Figure3C).

(A) EME, pH 10.0 (B) EPI, pH 3.0 (C) KET, pH 10.0

FigureFigure 3. 3.ChromatogramsChromatograms of of ( (AA)) emedastine (EME), (EME), ((BB)) epinastine epinastine (EPI), (EPI) and, and (C ()C ketotifen) ketotifen (KET) (KET) after afterUV UV/Vis/Vis irradiation irradiation in bu inﬀ bufferer solutions. solutions.

It is well-known that the photodegradation of drugs is significantly affected by the pH of the environment because the changes in pH can promote or inhibit photolysis [26]. In addition, the intermediates and final products that are generated in the wake of these photochemical reactions may also change the pH of the environment. Moreover, the drug may be more stable in its ionized or non-ionized form and may undergo specific acid-base catalysis in aqueous solutions [26,27]. The drugs analyzed in this study are all weak bases with a pKa value of 8.68 for EME (the substituted diazepine ring), 8.77 for EPI (due to primary amine group substituted in the imidazo [1,5-a]azepine structure), and 8.43 for KET (the substituted piperidine ring) [28-30] (Figure 1). However, they differ as far as their percentage degradation in specific pH is concerned. Alkaline conditions showed a higher degradation efficiency and the maximum degradation for EME and KET was obtained at pH 10.0, where they were electrically neutral. On the other hand, a different trend in the case of EPI was observed and the maximum degradation occurred at pH 3.0, where its primary amine group was protonated (Table 2).

Pharmaceutics 2020, 12, 560 8 of 22

It is well-known that the photodegradation of drugs is significantly affected by the pH of the environment because the changes in pH can promote or inhibit photolysis [26]. In addition, the intermediates and final products that are generated in the wake of these photochemical reactions may also change the pH of the environment. Moreover, the drug may be more stable in its ionized or non-ionized form and may undergo specific acid-base catalysis in aqueous solutions [26,27]. The drugs analyzed in this study are all weak bases with a pKa value of 8.68 for EME (the substituted diazepine ring), 8.77 for EPI (due to primary amine group substituted in the imidazo [1,5-a]azepine structure), and 8.43 for KET (the substituted piperidine ring) [28–30] (Figure1). However, they differ as far as their percentage degradation in specific pH is concerned. Alkaline conditions showed a higher degradation efficiency and the maximum degradation for EME and KET was obtained at pH 10.0, where they were electrically neutral. On the other hand, a different trend in the case of EPI was observed and the maximum degradation occurred at pH 3.0, where its primary amine group was protonated (Table2).

Table 2. Kinetics of the photodegradation of emedastine (EME), epinastine (EPI), and ketotifen (KET) in buﬀer solutions (n = 3).

1 Conditions Degradation (%) y = ax + b r K (min− ) t0.1 (h) t0.5 (h) EME Buffer pH 3.0 32.38 y = 0.0107x + 3.8306 0.9136 4.11 10 4 4.26 28.12 − × − Buffer pH 7.0 38.34 y = 0.0143x + 3.8604 0.9390 5.48 10 4 3.19 21.04 − × − Buffer pH 10.0 41.52 y = 0.0151x + 3.8371 0.9543 5.79 10 4 3.02 19.92 − × − EPI Buffer pH 3.0 19.10 y = 0.0054x + 3.8728 0.9351 2.07 10 4 8.44 55.72 − × − Buffer pH 7.0 12.16 y = 0.0044x + 3.9112 0.9860 1.69 10 4 10.4 68.39 − × − Buffer pH 10.0 8.10 y = 0.0028x + 3.9214 0.9604 1.07 10 4 16.3 107.5 − × − KET Buffer pH 3.0 14.64 y = 0.0046x + 3.9004 0.9823 1.77 10 4 9.91 65.42 − × − Buffer pH 7.0 19.28 y = 0.0049x + 3.8804 0.9303 8.87 10 4 1.97 13.03 − × − Buffer pH 10.0 100 - - - - -

3.4. Kinetics of Degradation Our study showed stronger correlations for the plots of logarithms of the concentration of non-degraded drugs than for the plots of the concentration of non-degraded drugs versus time of degradation, confirming pseudo first-order kinetics for their degradation, with all r values above 0.9. From these regression equations, further kinetic parameters were calculated, i.e., the degradation rate constant (k), degradation time of the 10% substance (t0.1), and degradation time of the 50% substance (t0.5) (Table2). As far as the photodegradation of EME is concerned, the calculated t0.5 values were 28.12 h at pH 3.0, 21.04 h at pH 7.0, and 19.92 h at pH 10.0, confirming its lowest stability in alkaline conditions. At the same time, EPI was shown to be less sensitive to UV/Vis light, with a percentage of degradation below 20%, even after the maximal dose of light. The calculated t0.5 values for EPI were 55.72 h at pH 3.0, 68.39 h at pH 7.0, and 107.5 h at pH 10.0, confirming its lowest stability in acidic conditions. As far as KET is concerned, the calculated t0.5 values dropped dramatically from 65.42 at pH 3.0 to 13.03 h at pH 7.0. What is more, the drug was completely degraded in pH 10.0, confirming its lowest stability in alkaline medium. The percentage degradation, as well as kinetic parameters for EME, EPI, and KET, are shown in Table2. In turn, the pseudo first-order profiles of degradation of EME, EPI, and KET in solutions of different pH values are shown in Figure4. Pharmaceutics 2020, 12, x FOR PEER REVIEW 9 of 24

Table 2. Kinetics of the photodegradation of emedastine (EME), epinastine (EPI), and ketotifen (KET) in buffer solutions (n = 3).

Degradation Conditions y = ax + b r K (min−1) t0.1 (h) t0.5 (h) (%) EME Buffer pH 3.0 32.38 y = −0.0107x + 3.8306 0.9136 4.11 × 10−4 4.26 28.12 Buffer pH 7.0 38.34 y = −0.0143x + 3.8604 0.9390 5.48 × 10−4 3.19 21.04 Buffer pH 10.0 41.52 y = −0.0151x + 3.8371 0.9543 5.79 × 10−4 3.02 19.92 EPI Buffer pH 3.0 19.10 y = −0.0054x + 3.8728 0.9351 2.07 × 10−4 8.44 55.72 Buffer pH 7.0 12.16 y = −0.0044x + 3.9112 0.9860 1.69 × 10−4 10.4 68.39 Buffer pH 10.0 8.10 y = −0.0028x + 3.9214 0.9604 1.07 × 10−4 16.3 107.5 KET Buffer pH 3.0 14.64 y = −0.0046x + 3.9004 0.9823 1.77 × 10−4 9.91 65.42 Buffer pH 7.0 19.28 y = −0.0049x + 3.8804 0.9303 8.87 × 10−4 1.97 13.03 Buffer pH 10.0 100 - - - - -

3.4. Kinetics of Degradation Our study showed stronger correlations for the plots of logarithms of the concentration of non-degraded drugs than for the plots of the concentration of non-degraded drugs versus time of degradation, confirming pseudo first-order kinetics for their degradation, with all r values above 0.9. From these regression equations, further kinetic parameters were calculated, i.e., the degradation rate constant (k), degradation time of the 10% substance (t0.1), and degradation time of the 50% substance (t0.5) (Table 2). As far as the photodegradation of EME is concerned, the calculated t0.5 values were 28.12 h at pH 3.0, 21.04 h at pH 7.0, and 19.92 h at pH 10.0, confirming its lowest stability in alkaline conditions. At the same time, EPI was shown to be less sensitive to UV/Vis light, with a percentage of degradation below 20%, even after the maximal dose of light. The calculated t0.5 values for EPI were 55.72 h at pH 3.0, 68.39 h at pH 7.0, and 107.5 h at pH 10.0, confirming its lowest stability in acidic conditions. As far as KET is concerned, the calculated t0.5 values dropped dramatically from 65.42 at pH 3.0 to 13.03 h at pH 7.0. What is more, the drug was completely degraded in pH 10.0, confirming its lowest stability in alkaline medium. The percentage degradation, as well as kinetic parameters for PharmaceuticsEME,2020 EPI, 12, ,and 560 KET, are shown in Table 2. In turn, the pseudo first-order profiles of degradation of 9 of 22 EME, EPI, and KET in solutions of different pH values are shown in Figure 4.

pH 3.0 4 3.9 3.8 3.7 3.6

3.5 KET LnC 3.4 EME 3.3 EPI 3.2 3.1 3 0 10 20 30 40 time (h) Pharmaceutics 2020, 12, x FOR PEER REVIEW 10 of 24

pH 7.0 4 3.9 3.8 3.7 3.6

3.5 KET LnC 3.4 EME 3.3 EPI 3.2 3.1 3 0 10 20 30 40 time (h)

pH 10.0 4 3.9 3.8 3.7 3.6

3.5 KET LnC 3.4 EME 3.3 EPI 3.2 3.1 3 0 10 20 30 40 time (h)

Figure 4.FigureFirst-order 4. First- plotsorder plots of the of degradationthe degradation of of emedastine emedastine (EME), (EME), epinastine epinastine (EPI) (EPI),, and ketotifen and ketotifen (KET) in(KET) buﬀer in solutions buffer solutions (n = 3,(n mean= 3, meanSD). ± SD). ± Because of the limited data from the literature concerning the photodegradation of EME, we Because of the limited data from the literature concerning the photodegradation of EME, we could could not conduct any detailed comparisons. As far as EPI is concerned, it was previously shown to not conductbe photostable any detailed in a solid comparisons. state [13], whereas As far a 2% as EPIdegradation is concerned, value was it observed was previously for a methanolic shown to be photostablesolution in a [12 solid]. However, state [ 13 no], results whereas on its a 2%degradation degradation in solutions value of was different observed pH values for a and methanolic its solutiondegradation [12]. However, kinetics no have results been published on its degradation so far. In terms in of solutions KET, only of one di similarﬀerent study pH valueshas been and its reported, in which KET was shown to be more stable than other benzocycloheptane antihistaminic agents [31]. In our study, KET was shown to be moderately sensitive to light at pH 3.0 and 7.0, and extremely sensitive at pH 10.0. Therefore, our results showed that EME and KET are sensitive to light in a wide pH range, which could be essential for manufacturing and protecting their formulations, especially in solutions. What is more, choosing the optimal excipients for their ocular drops becomes an important issue. As was described above, there are no articles in the literature concerning the kinetics of degradation with respect to EME, EPI, and KET. Therefore, the results presented here supplement the literary resources in this area.

3.5. LC/MS

Pharmaceutics 2020, 12, 560 10 of 22 degradation kinetics have been published so far. In terms of KET, only one similar study has been reported, in which KET was shown to be more stable than other benzocycloheptane antihistaminic agents [31]. In our study, KET was shown to be moderately sensitive to light at pH 3.0 and 7.0, and extremely sensitive at pH 10.0. Therefore, our results showed that EME and KET are sensitive to light in a wide pH range, which could be essential for manufacturing and protecting their formulations, especially in solutions. What is more, choosing the optimal excipients for their ocular drops becomes an important issue. As was described above, there are no articles in the literature concerning the kinetics of degradation with respect to EME, EPI, and KET. Therefore, the results presented here supplement the literary resources in this area.

3.5. LC/MS The identification of photodegradation products (Ps) of EME, EPI, and KET was performed on Pharmaceutics 2020, 12, x FOR PEER REVIEW 11 of 24 the basisPharmaceutics of our 20 UPLC20, 12, x/ FORMS PEER analysis REVIEW and supported with fragmentation patterns obtained11 of from 24 MS/MS + experiments.The identification A protonated of moleculephotodegradation [M + H] prodofuctsm/ z(Ps)303.2 of EME, was EPI observed, and KET in was the fullperformed scan mass on spectra of EMEthe(Table basis 3 of). our Its MS UPLC/MS/MS showed analysis the and product supported ion at withm/ z fragmentation246.2 that followed patterns a obtainedparallel fragmentation from + pathwayMS/MS to experiments. form ions at A mprotonated/z 232.1 and molecule 218.1. [M The + H] subsequent+ of m/z 303.2 was step observed in the fragmentationin the full scan mass pathway of EMEspectra was the of lossEME of (Table the water3). Its MS/MS moiety showed from m the/z 218.1, product which ion at resultedm/z 246.2 inthat the followed formation a parallel of a fragment fragmentation pathway to form ions at m/z 232.1 and 218.1. The subsequent step in the with mfragmentation/z 200.1. Following pathway this, to form the last ionsfragment at m/z 232.1 formed and an 218.1. ion at Them /z subsequent174.1 that step followed in the a parallel fragmentation pathway of EME was the loss of the water moiety from m/z 218.1, which resulted in fragmentation to form two ions at m/z 146.1 and 134.1. All of these steps in the fragmentation pathway the formation of a fragment with m/z 200.1. Following this, the last fragment formed an ion at m/z of EME174.1 are that shown followed in a Figureparallel5 fragmentationA. As was discussedto form two above,ions at m/z EME 146.1 was and 134.1. observed All of tothese be steps photolabile, with 32.38%,in the fragmentation 38.34%, and pathway 41.52% of degradationEME are shown at in pH Figure 3.0, 7.0,5A. As and was 10.0, discussed respectively. above, EME The was degradation processobserved at pH to 3.0 be was photolabile found to, with affect 32.38%, the ethoxy 38.34%, moiety, and 41.52% leading degradation to the productat pH 3.0, EME-P1 7.0, and with10.0, m/z 301.2 via hydroxylationrespectively. The and degradation dehydration, process while at pH 3.0 the was fragment found to ionsaffect formed the ethoxy from moiety, EME leading P-1 were to them /z 244.1, product EME-P1 with m/z 301.2 via hydroxylation and dehydration, while the fragment ions formed 218.1,product and 146.1 EME (Figure-P1 with5 m/zB). 301.2 In addition, via hydroxylation the degradation and dehydration at pH, 3.0, while 7.0, the and fragment 10.0 wasions shownformed to a ffect from EME P-1 were m/z 244.1, 218.1, and 146.1 (Figure 5B). In addition, the degradation at pH 3.0, 7.0, the 1,4-diazepine ring of EME, leading to its opening and to the product EME-P2 with m/z 277.2, with and 10.0 was shown to affect the 1,4-diazepine ring of EME, leading to its opening and to the subsequentproduct fragmentEME-P2 with ions m/z at277.2m/z, with246.2, subsequent 218.1, 200.1, fragment 174.1, ions 146.1,at m/z 246.2, and 218.1, 134.1 200.1, (Figure 174.1,5C). 146.1 It, is worth mentioningand 134.1 that (Figure the EME-P25C). It is worth product mentioning was described that the EME in European-P2 product Pharmacopoeia was described in asEuropean Impurity F [19]. A respectivePharmacopoeia UPLC as chromatogram Impurity F [19]. of A EMErespective and itsUPLC degradation chromatogram products of EME is and shown its degradation in Figure6 . products is shown in Figure 6. Table 3. Proposed structures of the photodegradation products of emedastine (EME). Table 3. Proposed structures of the photodegradation products of emedastine (EME). + Compound tRR (min) [M + H] + Fragmentation Ions Structure Compound tR (min) [M + H] + Fragmentation Ions Structure

134.1,134.1, 146.1, 146.1, 174.1, 174.1, 200.1, EMEEME 2.20 2.20 303.2 303.2 218.1,218.1, 232.1, 232.1, 246.2

EMEEME-P1-P1 pH pH 3.0 3.0 2.02 2.02 301.2 301.2 146.1, 146.1, 218.1, 218.1, 244.1

EMEEME-P2-P2 = Impurity= Impurity F [19 F] [ 19pH] 134.1,134.1, 146.1, 146.1, 174.1, 174.1, 200.1, 2.492.49 277.2 277.2 3.0pH pH 3.0 pH7.0 7.0pH pH10.0 10.0 218.1,218.1, 246.2 246.2

Pharmaceutics 2020, 12, 560 11 of 22 Pharmaceutics 2020, 12, x FOR PEER REVIEW 12 of 24

(A)

(B)

(C)

FigureFigure 5. Proposed 5. Proposed fragmentation fragmentation patterns patterns of of emedastineemedastine (EME) (EME) (A (A), ),EME EME-P1-P1 (B) (,B and), and EME EME-P2-P2 (C). (C).

Pharmaceutics 2020, 12, 560 12 of 22 Pharmaceutics 2020, 12, x FOR PEER REVIEW 13 of 24 Pharmaceutics 2020,, 12,, xx FORFOR PEERPEER REVIEWREVIEW 13 of 24

Figure 6. UPLC chromatogram of emedastine (EME) and its degradation products in a solution of pH 3.0. Figure 6.. UPLC chromatogram of emedastine (EME) and its degradation products in a solution of FigurepH 3.0.6. UPLC chromatogram of emedastine (EME) and its degradation products in a solution of As far as EPI is concerned, a protonated molecule [M + H]+ of m/z 250.1 was observed in the full pH 3.0. scan massAs far spectra as EPI (Table is concerned,4). Its MSa protonated/MS showed molecule the product[M + H]++ of ion m/z at 250.1m/z 208.1 was observed that followed iin the full a parallel scan mass spectra (Table 4). Its MS/MS showed the product ion at m/z 208.1 that followed a parallel fragmentationscanAs massfar as spectra EPI to is form (Tableconcerned, ions 4). atItsm aMS/MS/ zprotonated193.1 showed and molecule 178.1. the product Next, [M ionthe+ H at last]+ m/zof fragmentm/z 208.1 250.1 that formed wasfollowed observed ions a parallel at inm the/ z 130.1, full fragmentation to form ions at m/z 193.1 and 178.1. Next, the last fragment formed ions at m/z 130.1, scan115.1,fragmentation mass and spectra 91.1.The to (Table form subsequent ions 4). atIts m/z MS/MS steps 193.1 in showedand the fragmentation178.1. the Next, product the pathwaylast ion fragment at m/z of EPI, 208.1form includinged that ionsions followed atat them/z formation130.1, a parallel of other115.1 minor,, and fragments,91.1. The subsequent are depicted steps inin Figurethe fragmentation7. The photodegradation pathway of EPI,, ofincludingincluding EPI achieved the formation 8.10–19.10%, fragmentationof other minor to form fragme ionsnts at, arem/z depicted 193.1 and in 178.1. Figure Next, 7. The the plasthotodegradation fragment form ofed EPI ions achieved at m/z 130.1, whileof other the pH minor of the fragme buﬀernts decreased,, are depicted from in 10 Figure to 3. At7.. pHThe 3.0, photodegradation degradation was of EPI found achieved to aﬀ ect the 115.18.1, 0and–19 .10%91.1., whileThe subsequent the pH of the steps buffer in decreased the fragmentation from 10 to 3. pathway At pH 3.0 of, degradation EPI, including was foundthe formation to imidazoline8.10–19.10% ring,, while leading the pH to of its the oxidation, buffer decreased as was observedfrom 10 to for3. At the pH products 3.0, degradation EPI-P1, was EPI-P2, found and to EPI-P3. of otheraffect the minor imidazoline fragme ring,nts ,leading are depicted to its oxidation, in Figure as was 7 observed. The p forhotodegradation the products EPI of-P1, EPI EPI -P2 achieved,, The major ions formed during their MS/MS were m/z 248.1, 204.1, 178.1, and 127.1 (EPI-P1); 246.1, 8.10and–19 .10% EPI-P3., while The the major pH ions of the formed buffer during decreased their MS/MSfrom 10 were to 3. m/z At pH248.1, 3.0 204.1,, degradation 178.1,, and was 127.1 found to 204.1, and 178.1 (EPI-P2); and 204.1, 178.1, and 127.1 (EPI-P3). Further degradation of EPI led to affect(EPI the-P1) imidazoline;; 246.1, 204.1 ,ring,, and 178.1leading (EPI to-P2) its;; andoxidation, 204.1, 178.1 as was,, and observed 127.1 (EPI for-P3). the Further products degradation EPI-P1, of EPI -P2, andimidazoleEPI EPI led-P3. to ring imidazole The opening, major ring ions as opening, was formed observed as was during observed for thetheir productfor MS/MS the product EPI-P4 were EPI withm/z-P4 248.1, thewith fragment the 204.1, fragment 178.1 ions ions, at and mat/ z 127.1237.1, (EPI210.1,m/z-P1) and237.1,; 246.1, 192.1. 210.1 204.1,, It and was, and 192.1. interesting 178.1 It was (EPI interesting to-P2) observe; and to 204.1, thatobserve EPI-P3178.1 that, and EPI and -127.1P3 EPI-P4 and (EPI EPI were--P3).P4 were detectedFurther detected degradation both both at pH 3.0,of EPIwhereat led pH theto 3.0imidazole highest,, where degradation the ring highest opening, degradationof EPI as was occurred observed of EPI (19.10%), occurred for the as (19.10%) wellproduct as,, atas EPI pH well-P4 10.0, aswith at where pHthe 10.0,fragment degradation where ions was at degradation was relatively low (8.10%). At the same time, the EPI-P3 was described in European m/zrelativelyd 237.1,egradation low210.1 (8.10%).was, and relatively 192.1. At theIt low was same (8.10%). interesting time, At the the to EPI-P3 same observe time was ,,that the described EPIEPI--P3P3 was inand European described EPI-P4 were Pharmacopoeiain European detected both as Pharmacopoeia as Impurity A [19]. The fragmentation patterns of EPI-Ps 1–4 are shown in Figure atImpurity pHPharmacopoeia 3.0, Awhere [19]. as theThe Impurity highest fragmentation A degradation [19]. The patterns fragmentation of EPI of EPI-Ps occurred patterns 1–4 (19.10%) are of EPI shown-Ps, as1– in4 well are Figure shown as at8A–D, pHin Figure 10.0, while where their 8A–D,, while their UPLC chromatogram is shown in Figure 9.. dUPLCegradation chromatogram was relatively is shown low in(8.10%). Figure 9At. the same time, the EPI-P3 was described in European PharmacopoeiaTable as Impurity 4. Proposed A structures [19]. The of thefragmentation photodegradation patterns products of of EPI epinastine-Ps 1– (EPI).4 are shown in Figure Table 4. Proposed structures of the photodegradation products of epinastine (EPI). 8A–D, while their UPLC chromatogram is shown in Figure 9. tR Compound [M + H] ++ + Fragmentation Ions Structure Compound tR ((min)min) [M + H] Fragmentation Ions Structure Table 4. Proposed structures of the photodegradation products of epinastine (EPI).

tR Compound [M + H] + Fragmentation Ions Structure (min) 91.1,91.1, 115.1, 115.1, 130.1, 130.1, 165.1, 165.1, 178.1, EPIEPI 3.793.79 250.1 250.1 178.1,193.1, 193.1, 208.1 208.1

91.1, 115.1, 130.1, 165.1, 178.1, EPI 3.79 250.1 193.1, 208.1

EPI-P1 pH 3.0 3.30 266.1 127.1, 178.1, 204.1, 248.1 EPI-P1EPI-P1 pH pH 3.0 3.0 3.303.30 266.1 266.1 127.1, 127.1, 178.1, 178.1, 204.1, 204.1, 248.1 248.1

EPI-P1 pH 3.0 3.30 266.1 127.1, 178.1, 204.1, 248.1

EPI-P2EPI-P2 pH pH 3.0 3.0 3.433.43 264.1 264.1 178.1, 178.1, 204.1, 204.1, 246.1 246.1

EPI-P2 pH 3.0 3.43 264.1 178.1, 204.1, 246.1

Pharmaceutics 2020, 12, 560 13 of 22

Pharmaceutics 2020, 12, x FOR PEER REVIEW Table 4. Cont. 14 of 24 Pharmaceutics 2020, 12, x FOR PEER REVIEW 14 of 24 + Compound tR (min) [M + H] Fragmentation Ions Structure

EPI-P3 = Impurity A [19] pH 3.0 EPIEPI-P3-P3 = Impurity= Impurity A [ A19 []19 pH] 3.0 4.06 248.1 127.1, 178.1, 204.1 pH 10.0 4.064.06 248.1 248.1 127.1, 127.1, 178.1, 178.1, 204.1 204.1 pH 3.0pH pH 10.0 10.0

EPIEPI-P4EPI-P4- P4pH pH pH 3.0 3.0 3.0 pH pH pH 10.0 10.0 3.353.35 270.1270.1 270.1 192.1, 192.1,192.1, 210.1, 210.1, 210.1, 237.1 237.1 237.1

Figure 7. Proposed fragmentation pattern of epinastine (EPI).

Pharmaceutics 2020, 12, 560 14 of 22 Pharmaceutics 2020, 12, x FOR PEER REVIEW 15 of 24

(A)

(B)

(C)

(D)

FigureFigure 8. 8Proposed. Proposed fragmentation fragmentation patternspatterns ofof EPI-P1EPI-P1 ( A),), EPI EPI-P2-P2 ( (BB),), EPI EPI-P3-P3 (C (C), ),and and EPI EPI-P4-P4 (D ()D. ).

Pharmaceutics 2020, 12, x FOR PEER REVIEW 16 of 24

PharmaceuticsPharmaceutics 20202020, ,1122,, x 560 FOR PEER REVIEW 1615 of of 24 22

Figure 9. UPLC chromatogram of epinastine (EPI) and its degradation products in a solution of pH

Figure 9. UPLC chromatogram of epinastine3.0. (EPI) and its degradation products in a solution of pH 3.0. Figure 9. UPLC chromatogram of epinastine (EPI) and its degradation products in a solution of pH + LastlyLastly,, KETKET showed showed a protonated a protonated molecule molecule [M 3.0.+ [M H ]++ atH] m/zat 310.1m/z 310.1 in the in full the scan full scanmass mass spectra spectra,, whilewhile its MS/MS its MS /displayedMS displayed product product ions ionsat m/z at 249.0,m/z 249.0, 213.1 213.1,, and and96.1 96.1(Table (Table 5). The5). Thesubsequent subsequent steps steps in thein fragmentation theLast fragmentationly, KET showedpathway pathway a ofprotonated KET of KETare shown aremolecule shown in Figure [M in + Figure H 10]+ . atThe 10 m/z. drug The 310.1 drug was in wasthedegraded full degraded scan in mass14.64% in 14.64% spectra at at, pH 3.0whilepH and 3.0 its in and MS/MS 19.28% in 19.28% displayed at pH at 7.0. pH product What 7.0. is What ionsmore, isat almost more,m/z 249.0, almost100% 213.1 degradation 100%, and degradation 96.1 of(Table KET 5).was of The KET obse subsequent wasrved observed at pH steps at 10.0.inpH In the all 10.0. fragmentationpH In conditions all pH conditions pathway degradation of degradation KET was are found shown was to found affectin Figure tothe a ﬀ10piperidineect. The the drug piperidine ring, was leading degraded ring, to leading all in of14.64% tothe all at of identifiedpHthe 3.0 identiﬁed products, and in 19.28% products, i.e., KET at pH i.e.,-P1, 7.0. KET-P1, KET What-P2, KET-P2,is and more, KET andalmost-P3 KET-P3,, via 100% its oxidation viadegradation its oxidation.. Additionally, of KET Additionally, was oxidationobserved oxidation ofat pH the 10.0.methylof the In methyl allsubstituent pH substituentconditions to a formyl degradation to a formyl moiety moiety waswas foundobserved was observedto affect at pH the at 3.0 pH piperidine (KET 3.0 (KET-P2)-P2) andring, demethylationand leading demethylation to all of the of the nitrogenidentifiedthe nitrogen atom products, atom of the of i.e.piperidine the, KET piperidine-P1, moiety KET moiety- P2was, and wasobserved KET observed-P3 at, pHvia at 7.0its pH oxidation(KET 7.0 (KET-P3).-P3).. TheAdditionally, major The major ions oxidation ionsformed formed of duringtheduring theirmethyl their MS/MS substituent MS /wereMS were m/zto a 324.1,m formyl/z 324.1, 306.1, moiety 306.1, 278.1, was 278.1, and observed and221.0 221.0 (KET at pH (KET-P1);-P1) 3.0; 292.1, (KET 292.1, -263.2,P2) 263.2,and 275.1 demethylation 275.1,, and and221.0 221.0 of (KETthe(KET-P2);-P2) nitrogen; and and324.1, atom 324.1, 306.1 of the 306.1 278.1, piperidine 278.1, 250.1 250.1, ,moiety and and 221.0 221.0was (KET observed (KET-P3).-P3). atThe The pH respective respective7.0 (KET -fragmentationP3). fragmentation The major patternsions patterns formed and and duringcorresponding their MS/MS UPLC were chromatogram m/z 324.1, 306.1, are shown 278.1, in and Figures 221.0 111 1(KET––1414.. -P1); 292.1, 263.2, 275.1, and 221.0 (KET-P2); and 324.1, 306.1 278.1, 250.1, and 221.0 (KET-P3). The respective fragmentation patterns and correspondingTableTable 5. Proposed 5.UPLCProposed structureschromatogram structures of the of photodegradationare the shown photodegradation in Figures products products11– of14 ketotifen. of ketotifen (KET). (KET).

+ + CompoundCompound tR (min) [M t +(min) H] + [M + H]FragmentationFragmentation Ions IonsStructure Structure Table 5. ProposedR structures of the photodegradation products of ketotifen (KET).

Compound tR (min) [M + H] + Fragmentation Ions Structure

96.1, 185.0, 199.0, 213.1, KET KET3.79 310.1 3.79 96.1, 310.1 185.0, 199.0, 213.1, 221.0, 249.0 221.0, 249.0

KET 3.79 310.1 96.1, 185.0, 199.0, 213.1, 221.0, 249.0

KET-P1 pHKET-P1 7.0 pH3.43 7.0 342.1 3.43 342.1324.1, 306.1, 324.1, 278.1, 306.1, 221.0 278.1, 221.0

KET-P1 pH 7.0 3.43 342.1 324.1, 306.1, 278.1, 221.0

Pharmaceutics 2020, 12, 560 16 of 22

Table 5. Cont. Pharmaceutics 20202020,, 1122,, xx FORFOR PEERPEER REVIEWREVIEW 1717 ofof 2424 + Compound tR (min) [M + H] Fragmentation Ions Structure

Pharmaceutics 2020, 12, x FOR PEER REVIEW 17 of 24

KET--P2 pHKET-P2 7.0 pH3.56 7.0 320.1 3.56 320.1247.1, 275.1, 247.1, 263.1, 275.1, 292.1 263.1, 292.1

KET-P2 pH 7.0 3.56 320.1 247.1, 275.1, 263.1, 292.1

KET--P3 221.0, 250.1, 278.1, KET-P3KET-P3 pH 3.69 7.0 342.1 3.69 221.0, 342.1 250.1, 278.1, 306.1, 324.1 pH 7.0 3.69 342.1 221.0, 250.1, 278.1,306.1, 306.1, 324.1 324.1 pH 7.0

FigureFigure 10. 10.Proposed Proposed fragmentationfragmentation pattern pattern of ofketotifen ketotifen (KET). (KET).

Figure 10. Proposed fragmentation pattern of ketotifen (KET).

Pharmaceutics 2020, 12, x FOR PEER REVIEW 18 of 24 Pharmaceutics 2020, 12, 560 17 of 22

Pharmaceutics 2020, 12, x FORFigure PEER REVIEW 11. Proposed fragmentation pattern of KET-P1. 19 of 24

Figure 11. Proposed fragmentation pattern of KET-P1.

Figure 12. Proposed fragmentation pattern of KET-P2. Figure 12. Proposed fragmentation pattern of KET-P2.

Pharmaceutics 2020, 12, 560 18 of 22 PharmaceuticsPharmaceutics 2020, 12 20, x20 FOR, 12, PEERx FOR REVIEW PEER REVIEW 20 of 2420 of 24

FigureFigureFigure 13. Proposed 13. 13.Proposed Proposed fragmentation fragmentationfragmentation pattern pattern of pattern KET of-P3. ofKET KET-P3.-P3.

Figure 14. UPLC chromatogram of ketotifen (KET) and its degradation products in a solution of pH 7.0. FigureFigure 14. UPLC 14. UPLCchromatogram chromatogram of ketotifen of ketotifen (KET) and(KET) its and degradation its degradation products products in a solution in a solution of pH of pH 3.6. ROS Assays 7.0. 7.0. In the present experiments, the samples of EME, EPI, and KET were irradiated for 15, 30, and 3.6. ROS Assays 3.6.45 min,ROS Assays and for 1, 3, and 7 h, which corresponded to energies equal to 675, 1350, 2025, 2700, 8100, andIn 18,902 the Inpresent kJthe/m present2 experiments,. Bearing experiments, in the mind samples the that samples of the EME, indicator of EME,EPI, and EPI setting KET, and were KET value irradiatedwere of 250irradiated Wfor/ m15,2 for 30for ,15, and 1 30 h 45 achieved, and 45 an min,intensity andmin, for ofand 1, UVA 3for, and 1, from 3 7, andh 6.5, which 7 to h, 7.9 which corresponded J/cm corresponded2 for our to CPSenergies to Plus energies equal chamber, equalto 675, theto 1350, 675, selected 2025,1350, 2700, light2025, 8100, energies2700, and8100, comprised and 2 2 18,902the range18,902 kJ/m of2 .kJ/m Bearing 5–20. JBearing/cm in 2 mind, which in mindthat is the recommendedthat indicator the indicator setting in all setting value phototoxicity value of 250 of W/m 250 assays 2 W/mfor [ 120 for h, 22 achiev 1,23 h]. achieved To an bettered an verify intensity of UVA from 6.5 to 7.92 J/cm2 for our CPS Plus chamber, the selected light energies intensityour results, of UVA quinine from and 6.5 benzocaine to 7.9 J/cm were for included our CPS inPlus our chamber, ROS experiments the selected as light a positive energies and negative comprisedcomprised the range the rangeof 5–20 of J/cm 5–202, J/cmwhich2, which is recommended is recommended in all phototoxicityin all phototoxicity assays assays [20,22,23 [20,22,23]. To ]. To control, respectively. better betverifyter verify our results, our results, quinine quinine and benzocaine and benzocaine were includedwere included in our in ROS our experiments ROS experiments as a positive as a positive and negativeAsand a negative starting control, control, point respectively. to respectively. evaluate the phototoxic potential of EME, EPI, and KET, the photosensitized lipid peroxidation was monitored using linoleic acid as a substrate. It is known from the literature that phototoxic drugs accelerate the peroxidation of lipids upon exposure to UV/Vis radiation [22].

In addition, lipid peroxidation in the cellular membrane is considered to be one of the major mechanisms in drug-induced photoirritation [22]. Pharmaceutics 2020, 12, 560 19 of 22

In the present study, lipid peroxidation was monitored as TBARs determination with MDA as a standard. The linearity between the level of MDA and the absorbance at 532 nm due to the product of MDA with TBA was investigated. The regression line equation calculated by the least squares method was y = 1.1287 0.0056x + 0.0261 0.0011, with a coeﬃcient of correlation of r = 0.9997 0.0005 (mean ± ± ± SD, n = 5). On the basis of the above calibration equation, the number of TBARs generated under the ± irradiation of linoleic acid in the absence or presence of EME, EPI, and KET was determined. It was clearly shown that the peroxidation of linoleic acid was at the level of the control group when the drugs were absent. However, irradiation of the substrate in the presence of EME and KET produced a much larger number of TBARs. What is more, the results obtained for KET were similar to those of quinine, which is known to be a highly photoreactive substance. In contrast, the addition of EPI to linoleic acid increased peroxidation to a lesser extent, although the content of TBARs was higher than in the presence of benzocaine (a negative control) (Table6). At the same time, the obtained results were correlated with high values of MEC for EME and KET at wavelengths near 290 nm. Therefore, EME and especially KET can be considered to have a potent photoirritancy in vitro, on the basis of their ability to enhance the peroxidation of linoleic acid [22].

Table 6. Results of the ROS tests of emedastine (EME), epinastine (EPI), and ketotifen (KET) versus quinine as a positive control and benzocaine as a negative control (n = 3).

Energy (kJ/m2) EME EPI KET Quinine Benzocaine TBARs 675 0.107 0.087 0.120 0.162 0.055 1350 0.118 0.089 0.121 0.170 0.057 2025 0.122 0.091 0.136 0.175 0.062 2700 0.131 0.093 0.141 0.189 0.063 8100 0.133 0.101 0.207 0.244 0.078 18,902 0.136 0.109 0.274 0.285 0.081 SO 675 8 6 9 282 9 − 1350 10 8 11 319 8 − 2025 21 8 27 401 3 − 2700 37 23 55 493 10 8100 72 27 173 543 13 18,902 101 35 452 586 21 SA 675 29 6 60 133 11 − 1350 34 21 86 148 8 − 2025 49 29 110 179 6 − 2700 110 34 151 221 2 8100 216 44 358 422 8 18,902 226 62 424 540 9 TBARs, thiobarbituric acid reactive substances test; SO, singlet oxygen test; SA, superoxide anion test.

According to the official guidelines [20,22], the phototoxic potential of drugs can be elucidated by monitoring ROS generation under exposure to UV/Vis radiation, as SO (type 2 photochemical reaction) and SA (type 1 photochemical reaction) generation. In our experiments, each drug was judged to be photoreactive when (a) SO values of 25 or more and SA values of 70 or more were measured; (b) SO values of less than 25, but SA values of 70 or more, were measured; and (c) SO values of 25 or more and SA values of less than 70 were measured. When SO values were lower than 25 and SA values were between 20 and 70, the drug was judged as weakly photoreactive [22]. According to 2 2 these criteria and the standard intensity of UVA from 6.5 to 7.9 J/cm , corresponding to 2700 kJ/m , EME and KET were identified as phototoxic drugs by mechanisms including ROS generation (SO > 25 Pharmaceutics 2020, 12, 560 20 of 22 and SA > 70). In contrast, EPI was judged as weakly photoreactive (SO < 25 and 20 > SA < 70) (Table6). Therefore, EME and KET would have substantial photoreactivity and could potentially cause phototoxic reactions in vivo, upon exposure to sunlight. The obtained results clearly show the necessity for further experiments concerning the phototoxicity of EME and KET, especially considering that they are more vulnerable to degradation when applied topically [8]. Almost all antihistamines have been reported in the literature as causing hypersensitivity reactions, including their topical formulations. Shakouri and Bahna [32] published a comprehensive review on antihistaminic allergies reported between 1949 and 2013. More than one hundred cases were identified, which included urticaria, angioedema, anaphylaxis, fixed drug eruptions, allergic contact, and photosensitive dermatitis, as well as generalized nonspecific rashes [32,33]. Therefore, our observations in vitro could be in agreement with the phototoxicity data from these clinical studies [33]. Among different types of products, the stability of ophthalmic pharmaceuticals is the most important, since eyes are very sensitive. They are easily irritated if the composition of the ophthalmic pharmaceutical is not adequate [34]. Therefore, ophthalmic pharmaceuticals must be suitably compounded and packaged for their application to the eye [5]. The pH value is often the stability-controlling factor for many products, but it must be within a certain range because of the bioavailability of the drug [34]. Our results clearly show the need for optimal pH selection for EME and KET, bearing in mind both of the questions presented above.

4. Conclusions The present study was designed and performed considering the importance of EME, EPI, and KET in therapy and their topical use as ocular drops, as well as the lack of reports concerning their degradation, isolation, and characterization of their degradation products. As a result, we presented complete photostability data for EME, EPI, and KET at diﬀerent pH values, their kinetics of degradation, and characterization of their main photodegradation products. Several degradation products of EME, EPI, and KET were detected and identiﬁed by our UPLC-MS method that have not previously been described in the literature, i.e., EME-P1; EPI-Ps 1, 2, and 4; and KET-Ps 1–3. In addition, EME and KET were deduced to have a phototoxic risk in UV/Vis absorption measurements and their MEC values. Because of several false negative predictions based on the MEC values which were reported in the literature, a combined use of other photosafety testing is recommended to ensure the photosafety of drugs [20,21]. As a consequence, the MEC evaluation, ROS assays, and photosensitized peroxidation of linoleic acid were reported. Therefore, the results developed here can be utilized to improve the quality of topical formulations of EME, EPI, and KET. They can also be used to minimize the risk associated with their photodegradation, as well as the phototoxic risk of EME and KET in terms of patients’ health.

Author Contributions: Conceptualization, A.G. and A.B.-R.; methodology, A.G., A.B.-R., P.Z.,˙ K.L., and P.K.; software, A.G., P.Z.,˙ and P.K.; validation, A.G., K.L., and P.K.; formal analysis, A.G., A.B.-R., and U.H.; investigation, A.B.-R., K.L., P.K., and P.Z.;˙ resources, A.G. and A.B.-R.; data curation, A.G., A.B.-R., P.Z.,˙ K.L., and P.K.; writing—original draft preparation, A.G., U.H., and P.Z.;˙ writing—review and editing, A.G. and U.H.; supervision, U.H. All authors approved the final manuscript for publication. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflicts of interest.

References

1. Ahmad, I.; Ahmed, S.; Anwar, Z.; Ali Sheraz, M.; Sikorski, M. Photostability and photostabilization of drugs and drug products. Int. J. Photoenergy 2016, 7, 1–19. [CrossRef] 2. Bajaj, S.; Singla, D.; Sakhu, N. Stability testing of pharmaceutical Products. J. Appl. Pharm. Sci. 2012, 2, 129–138. Pharmaceutics 2020, 12, 560 21 of 22

3. Henry, B.; Foti, C.; Alsante, K. Can light absorption and photostability data be used to assess the photosafety risks in patients for a new drug molecule. J. Photochem. Photobiol. B Biol. 2009, 96, 57–62. [CrossRef] [PubMed] 4. Sanchez-Borges, M.; Ansotegui, J. Second generation antihistamines: An update. Curr. Opin. Allergy Immunol. 2019, 19, 358–364. [CrossRef] 5. Ben-Eli, H.; Solomon, A. Topical antihistamines, mast cell stabilizers, and dual-action agents in ocular allergy: Current trends. Curr. Opin. Allergy Immunol. 2018, 18, 411–416. [CrossRef] 6. Zgadzaj, A.; Kornacka, J.; Jastrz˛ebska,A.; Parzonko, A.; Sommer, S.; Nał˛ecz-Jawecki,G. Development of photoprotective, antiphototoxic and antiphotogenotoxic formulations of ocular drugs with ﬂuoroquinolones. J. Photochem. Photobiol. B Biol. 2018, 178, 201–210. [CrossRef][PubMed] 7. Jamrógiewicz, M. Consequences of new approach to chemical stability tests to active pharmaceutical ingredients. Front. Pharmacol. 2016, 7, 17. [CrossRef] 8. Kryczyk-Poprawa, A.; Kwiecie´n,A.; Opoka, W. Photostability of topical agents applied to the skin: A review. Pharmaceutics 2020, 12, 10. [CrossRef] 9. Salama, N.N.; Abdel-Razeq, S.A.; Abdel-Atty, S.; El-Kosy, N. Development and validation of densitometry TLC stability indicating method for quantitative determination of azelastine hydrochloride and emedastine hydrochloride in their drug products. Br. J. Pharm. Res. 2014, 4, 79–92. [CrossRef] 10. Dal Molim, D.; Steppe, M.; Schapoval, E. Development and validation of liquid chromatographic and ultraviolet derivative spectrophotometric methods for determination of epinastine hydrochloride in coated tablets. J. AOAC Int. 2007, 5, 1266–1271. 11. El-Bagary, R.; Boshra, A.; El-Hakeem, M.; Abdelrao, A. Three analytical methods for determination of epinastine hydrochloride in bulk and in ophthalmic solutions. J. Chem. Pharm. Res. 2012, 4, 1361–1369. 12. Malakar, A.; Bokshi, B. Development and validation of stability indicating RP-HPLC method for the determination of epinastine hydrochloride in pharmaceutical dosage form. Int. Curr. Pharm. J. 2012, 1, 50–55. [CrossRef] 13. Ahhirao, V.; Pawar, R. Stability-indicating LC method for the determination of epinastine in bulk drug and in pharmaceutical dosage form. Res. J. Recent Sci. 2012, 1, 281–288. 14. Baburao Ubale, M. A validated stability-indicating HPLC assay method for epinastine HCl in bulk drug. J. Curr. Chem. Pharmaceut. Sci. 2012, 2, 107–112. 15. Troche Concepción, Y.; Romero Diaz, J.A.; García Peña, C.M. Validation of analytical method to quality control and the stability study of 0.025% eye drops ketotifen. Rev. Cuba Farm. 2010, 44, 318–325. 16. Nnane, I.P.; Damani, L.A.; Hutt, A.J. Development and validation of stability indicating high-performance liquid chromatographic assays for ketotifen in aqueous and silicon oil formulations. Chromatographia 1998, 48, 797–802. [CrossRef] 17. Kabra, P.V.; Nargund, L.V.G.; Murthy, M.S. Development and validation of a stability-indicating LC-UV method for simultaneous determination of ketotifen and cetirizine in pharmaceutical dosage forms. Trop. J. Pharm. 2014, 13, 1141–1147. [CrossRef] 18. Abd El-Bary, A.A.; Salem, H.F.; Kharshoum, R.M. 2-Hydroxypropyl-β-cyclodextrin complex with ketotifen fumarate for eye drops preparations. Int. J. Drug Deliv. 2011, 3, 228–240. 19. European Directorate for the Quality of Medicines & HealthCare. European Pharmacopoeia, 9th ed.; European Pharmacopoeia, Council of Europe: Strasbourg, France, 2016. 20. Organization for Economic Co-Operation and Development. Guidelines for the Testing of Chemicals; Test Guideline No 495: Reactive Oxygen Species (ROS) Assay for Photoreactivity; OECD Publishing: Paris, France, 2019. 21. ICH Harmonised Tripartite Guideline Q1B: Stability Testing: Photostability Testing of New Drug Substances and Products; EMEA: London, UK, 1996. 22. Seto, Y.; Inoue, R.; Kato, M.; Yamada, S.; Onoue, S. Photosafety assessments on pirfenidone: Photochemical, photobiological and pharmacokinetic characterization. J. Photochem. Photobiol. B Biol. 2013, 120, 44–51. [CrossRef][PubMed] 23. Onoue, S.; Suzuki, G.; Kato, M.; Hirota, M.; Nishida, H.; Kitagaki, M.; Kouzuki, H.; Yamada, S. Non-animal photosafety assessment approaches for cosmetics based on the photochemical and photobiochemical properties. Toxicol. In Vitro 2013, 27, 2316–2324. [CrossRef][PubMed] Pharmaceutics 2020, 12, 560 22 of 22

24. International Conference on Harmonization. ICH Guideline S10: Guidance on Photosafety Evaluation of Pharmaceuticals; International Conference on Harmonization: EMA: London, UK, 2014. 25. International Conference on Harmonization. ICH Harmonized Tripartite Guideline Q2(R1): Validation of Analytical Procedures: Methodology; International Conference on Harmonization: EMEA: London, UK, 1995. 26. Ahmad, I.; Bano, R.; AliSheraz, M.; Ahmed, S.; Mirza, T. Photodegradation of levoﬂoxacin in aqeous and organic solvents: A kinetic study. Acta Pharm. 2013, 63, 223–229. [CrossRef] 27. Mohamed, A.; Salama, A.; Nasser, W.; Uheida, A. Photodegradation of ibuprofen, cetirizine and naproxen by

PAN-MWCNT/TiO2-NH2 nanoﬁber membrane under UV light irradiation. Environ. Sci. Eur. 2018, 30, 47. [CrossRef][PubMed] 28. Emedastine. Available online: https://www.drugbank.ca/drugs/DB01084 (accessed on 18 May 2020). 29. Epinastine. Available online: https://www.drugbank.ca/drugs/DB00751 (accessed on 18 May 2020). 30. Ketotifen. Available online: https://www.drugbank.ca/drugs/DB00920 (accessed on 18 May 2020). 31. Abounassif, M.A.; El-Obeid, H.A.; Gadkariem, E.A. Stability studies on some benzocycloheptane antihistaminic agents. J. Pharm. Biomed. Anal. 2005, 36, 1011–1018. [CrossRef][PubMed] 32. Shakouri, A.A.; Bahna, S.L. Hypersensitivity to antihistamines. Allergy Asthma Proc. 2013, 34, 488–496. [CrossRef] 33. Rutkowski, K.; Li, P.H.; Wagner, A. The paradox of antihistamine hypersensitivity. J. Allergy Clin. Immunol. 2018, 6, 258–259. [CrossRef][PubMed] 34. Baranowski, P.; Karolewicz, B.; Gajda, M.; Pluta, J. Ophtalmic drug dosage forms: Characterization and research methods. Sci. World J. 2014, 1–14. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).