Free radical induced degradation and computational studies of hydroxychloroquine in aqueous solution

Abstract

Hydroxychloroquine (HCQ) is a potential drug molecule for treating malaria. Recently it has also been tried as adjustment in Covid 19 therapy. Interaction of HCQ with free radicals is very important, which controls its stability in the environment where free radicals are generated unintentionally. In this report, we present detailed investigation on the reactions of hydrated electrons (e_aq⁻) and hydroxyl radical (^•OH) with HCQ in aqueous solution through electron pulse radiolysis technique and computational studies. The degradation of HCQ was found to be faster in the case of reaction with ^•OH radicals. However, the degradation could be substantially slowed down in the presence of antioxidants like ascorbic acid and gallic acid. This revealed that the stability of HCQ could be enhanced in an oxidative environment in the presence of these two compounds, which are easily available through food supplements. Various global and local reactivity parameters are also determined to understand the reactivity trend using Hard-Soft Acid-Base (HSAB) principle in the realm of the DFT methods. Computational studies were performed to elucidate the site-specific reactivity trend towards the electrophilic and nucleophilic attack by calculating the condensed Fukui index for various species of HCQ.

Graphical abstract

1. Introduction

Various drugs based on the derivatives of 4-aminoquinolone are widely used for the treatment of several diseases including rheumatology and dermatology [Scherbel et al., 1958; Yao et al., 2019; Sardana et al., 2020]. One such widely used drug is chloroquine (CQ), which is used against the disease malaria. Hydroxychloroquine (HCQ), contains a terminal OH group, which imparts a greater aqueous solubility due to hydrogen bonding and also widely used as antimalarial prevention drug and to treat diseases such as rheumatoid arthritis, and systemic lupus erythematosus [Tanenbaum and Tuffanelli, 1980; Zvi et al., 2012; Bell, 1983; Rynes, 1988; Dorner, 2010]. Recently, national and international medical organizations have recommended the treatment of coronavirus (COVID-19) with administration of HCQ in patients under certain conditions [Lei et al., 2020; Pastick et al., 2020].

It is a water-soluble compound having three pK_a values (<4, 8.3 and 9.7) similar to its analogue i.e. choloroquine indicating that these pK_a values are due to the nitrogen groups in the molecule [Schroeder and Gerber, 2014]. In order to get its optimal effect for the treatment of a disease, its stability in the biological environment also needs to be optimized for a better availability towards the disease treatment. In the biological environment under stress conditions, oxidative radicals like ^•OH are generated. These radicals are extremely reactive and known to degrade or decompose various organic compounds in aqueous system [Homlok and TakacsWojnarovits, 2013; Shinde et al., 2014; Correa et al., 2020; Wang and Xu, 2012]. Therefore, under stress conditions, HCQ may also be attacked by such free radicals and undergo degradation, which could lead to its lower stability and less effectiveness.

In this regards, Saini and Bansal have reported the degradation of HCQ via hydrolysis, oxidation, dry heat and photolysis processes [Saini and Bansal, 2013]. These authors have also reported different degradation products characterized by LC-MS-TOF and LC-PDA studies. However, a systematic investigation on its degradation pathway through reactions with free radicals like ^•OH and e_aq ⁻ are yet to be studied.

HCQ has three nitrogen containing basic functional groups with pK_a values of <4.0, 8.3 and 9.7, two of which would be protonated at neutral pH, very similar to chloroquine (CQ) molecule which has three pK_a values of <4.0, 8.4 and 10.2 [Warhurst et al., 2003]. Thus, the difference between the two compounds is the presence of the additional hydroxyl group on HCQ which imparts a greater aqueous solubility due to hydrogen bonding of the terminal OH groups. This hydrogen bonding also causes a lowered pK_a1 value of 9.7 in HCQ as compared to pK_a1 value of 10.2 in CQ. This trend is similar to the drop of pK_a value from triethylamine (pK_a = 10.74) to 2-diethylamino ethanol (pK_a = 9.7) [Little et al., 1990]. Using CQ as reference compound for the ionization constants, pK_a2 could be clearly identified with the aromatic amidinium ion by comparison with the parent substance, 4-amino-7-chloroquinoline (pK_a = 8.23) [Perrin, 1965]. Hence, the pK_a1 value of 9.7 is due to tertiary nitrogen containing ethanolic group. Similarly, the pK_a2 value of 8.3 is due to the aromatic nitrogen in the quinoline ring and the pK_a3 value of <4.0 is due to secondary amino group attached to the ring carbon atom in the 4th position.

In this report, we have investigated its degradation studies through the reaction with ^•OH and e_aq ⁻ as well as O₂ ^•- in aqueous solution at natural pH ∼6 through 2μs electron pulse irradiations (steady-state) as well as time-resolved ns electron pulse radiolysis studies. Various naturally available compounds such as ascorbic acid, oxalic acid, citric acid and gallic acid have been used for their effects on its oxidative degradation. Additionally, computational studies were also carried out to determine various reactivity parameters for different forms of HCQ. Site specific reactivity for nucleophilic and electrophilic addition reactions were predicted using various computational methods.

2. Methodology

Experimental: High purity (99%) hydroxychloroquine sulfate (HCQ) was generously supplied by Ipca pharmaceutical company, Mumbai. Ascorbic acid (99%), citric acid (99.5%), oxalic acid (98%), gallic acid (97%) and tert-butanol (99.5%) were purchased from sigma aldrich and were used without further purification. Nanopure water obtained from a Millipore water purifying unit was used for preparing the solutions. Freshly prepared solutions were used in the experiments.

7 MeV linear electron accelerator (LINAC) coupled with a kinetic spectrometer was used for the irradiation as well as pulse radiolysis studies. Samples were irradiated with electron pulses of FWHM 2 μs and absorbed dose 100 Gy/pulse for the investigation of degradation studies. 200 ns electron pulses were used for the transient absorption measurements in the case of electron pulse radiolysis studies. Absorbed dose was determined by using a chemical dosimeter, 0.01 M KSCN aerated solution, where Gε(SCN)₂ ^•- value at 475 nm was taken as 2.6 x 10⁻⁴ m² J⁻¹ [Buxton and Stuart, 1995]. The degradation studies were performed by monitoring the absorption spectra of the HCQ solution before and after irradiation of each absorbed doses using a Shimadzu 650 absorption spectrophotometer. The limit of detection for the spectrometer was 5 in the absorbance scale. The experiments were performed several times (at least three times) in order to get a reproducible result. High purity N₂, N₂O or O₂ gas was purged in the solution just before the experiments.

Computational: The most stable geometry of HCQ (see Scheme 1 ) has been optimized using ab initio methods employing Density Functional Theory (DFT) [Parr and Yang, 1995]. A number of low energy conformers may exist for HCQ molecule due to its large side-chain. To search the minimum energy conformer, various initial input structures of the different conformers were first optimized using semiempirical methods, such as PM6 theory. Subsequently, the most stable geometry is finally optimized at higher level of theory using Becke three-parameter hybrid functional with nonlocal correlation provided by the Lee−Yang−Parr (B3LYP) method is used utilizing Dunning's correlation-consistent double-ζ basis set, namely, cc-pVDZ. All the calculations are performed incorporating the solvent effect introduced using the polarizable continuum model (PCM) incorporating the integral equation formalism (IEF) with water as solvent. When PCM was used for the solvent effect, the error in the total polarization charge was well below the acceptable limit of 0.05. In the present study the value was ∼0.01. Optimized structures, namely, minima, were confirmed by frequency calculations at same level of and also provides the zero-point energy (ZPE) correction. For obtaining the theoretical absorption spectra, vertical transition energies were estimated for various singlet excited states using Time-Dependent Density Functional Theory (TD-DFT) employing B3LYP method. For calculation of excitation energies, a higher basis set, namely, 6–311+g(d) was used.

Scheme 1.

Optimized structure of neutral HCQ with atom numbering showing various pK_a values and also the H-bonding distances in Å.

Various global and local chemical reactivity parameters can be also evaluated in terms of Hard-Soft Acid-Base (HSAB) principle using the DFT methods [Pearson, 1997; Pearson, 2005; Chandrakumar and Pal, 2002]. In this context, global reactivity parameters (GRP) such as chemical potential (μ), chemical hardness (η), chemical softness (S) and local reactivity parameters (LRP) such as Fukui function or indices are evaluated for HCQ using the DFT methods. While the GRP gives an overall stability of the molecular system, the LRP indicates site-selectivity of a chemical system [Sánchez-Márquez, 2019]. Fukui functions are defined as the functional derivative of the chemical potential respect to the external potential produced by the nuclei at a constant electron number. Since the chemical potential is defined as the derivative of the density functional with respect to the electron density, Fukui functions are also defined as the derivative of the electron density with respect to the number of electrons at a constant potential. The Fukui function represents the changes in the electron density of a point (atom or functional group) with the number of electrons added or removed from the molecular system at constant potential. Fukui indices are, in short, reactivity indices which give us information about which atoms in a molecule have a larger tendency to either loose or accept an electron, which can be directly related to an electrophilic or a nucleophilic attack [Domingo et al., 2013; Sánchez-Márquez, 2019]. To perform these calculations, the finite differences method is employed and the Fukui index condensed to atom is obtained which is known as condensed Fukui function (CFF). The details of these calculations can be seen in the recent publication [Upadhyaya, 2022]. In the present study, natural population analysis (NPA) is carried out at B3LYP/cc-pZDZ level of theory including solvation. It should be noted that the negative indices are meaningless and should be ignored in arriving any conclusion. All of the calculations were performed using mainly the GAMESS quantum-chemistry suite and the Gaussian 09 package of programs [Schmidt et al., 1993; Frisch et al., 2010].

3. Results and discussions

Hydroxychloroquine sulfate (HCQ) is a water-soluble drug molecule which exhibits two strong absorption peaks at 330 and 342 nm (Fig. 1 ) with a molar extinction coefficient value of 17,610 mol⁻¹dm³cm⁻¹ at 342 nm at natural pH 6. However, it also exhibits pH dependent absorption spectra. The two strong absorption peaks with a higher absorbance ratio of 342 nm: 330 nm appear in the case of pH 3, 6 and 7. However, at pH 8, these two peaks appear with a lower absorbance ratio of 342 nm: 330 nm. At higher pH, 9 and 10, the absorption peak at 342 nm disappears with only one absorption peak at 330 nm. All the radiation chemical experiments were carried out at natural pH 6, where it is expected to exist in the doubly protonated form. Due to high molar extinction coefficient, the concentration of HCQ was kept at 1 x 10⁻⁴ mol dm⁻³ with an absorbance value at 342 nm close to 2.0, in the solution throughout the study.

Fig. 1.

Absorption spectra of aqueous solution containing 0.1 mM HCQ at different pHs.

The computational method is used in the present study for evaluating the absorption spectra for the HCQ system. At intermediate or the experimental pH = 6.0, the HCQ molecule may exists in various protonated forms. However, at pH = 10, the HCQ exists predominantly as neutral form. The experimental absorption spectra at pH = 10.0 is identical to the absorption spectra of neutral CQ. Hence, the absorption spectra at pH = 10 is considered for the comparison with the computed absorption spectra for neutral HCQ molecule. Fig. 2 I shows such comparison. It can be seen that there is an excellent match between the experimental and computed absorption spectra considering the level of theory and the size of the molecule. The analysis of the molecular orbitals (MOs) involved in the electronic transition shows that the first absorption band is solely due to HOMO-LUMO transition. This band is π-π* in nature associated with oscillator strength value of 0.1844. The analysis also shows that all the electronic transitions are basically due to the quinoline ring only. The absorption spectra of monoprotonated and diprotonated HCQ species are also computed at same level of theory. The computed absorption spectra of monoprotonated, diprotonated along with the neutral HCQ molecule is shown in Fig. 2(II). The monoprotonated species shows similar absorption spectra as compared to the neutral HCQ molecule while the diprotonated species shows a blue shifted first absorption peak. For second and third absorption peaks, the diprotonated species shows red shifted peaks. The experimental spectra at pH∼6.0 shows three peaks in the vicinity of 330 nm, first shoulder peak at ∼317 nm, second peak at ∼330 nm and third peak at ∼342 nm. Similar peak structures are seen in the vicinity of ∼255 nm. Hence, it is assumed that these peaks are the vibronic structure of H₂HCQ⁺⁺ spectra which is blue shifted to the peak of neutral species as predicted by computational methods. Considering these aspects of the computed spectra, it is assumed that at pH = 6.0, HCQ exists predominantly as di-protonated species.

Fig. 2.

(I) (A) Absorption spectra of HCQ molecule at pH = 10 in which it exists predominantly as neutral form. (B) Theoretically calculated absorption spectra for neutral HCQ in water. (II) Theoretically calculated absorption spectra for various species of HCQ in water.

3.1. Pulse radiolysis studies

Radiolysis of water is known to yield various free radicals as well as molecular products [Caër, 2011; Spinks and Woods, 1976].

$${\mathrm{H}}_2\mathrm{O}{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-},\to {}^{•}\mathrm{O}\mathrm{H},{\mathrm{H}}^{•},{\mathrm{H}\mathrm{O}}_2^{•},{\mathrm{H}}_3{\mathrm{O}}^{+},{\mathrm{H}}_2,{\mathrm{H}}_2{\mathrm{O}}_2$$ (1)

The kinetics and dynamics of the reactions of primary free radicals like e_aq ⁻ and ^•OH radicals with HCQ were investigated by electron pulse radiolysis studies using a 7 MeV LINAC coupled with kinetic spectrometer. Electron pulses with FWHM 100 ns and absorbed dose of 15 Gy/pulse were used for this study. Each sample was irradiated only with a single electron pulse during the acquisition of transient absorption spectra. N₂O-saturated HCQ solution was used for the study of ^•OH radical reaction, whereas N₂ purged solution containing HCQ along with t-BuOH was used to study the reaction of e_aq ⁻ reaction with HCQ.

Fig. 3a shows the transient absorption spectra of the intermediate species formed in the reaction of ^•OH radicals with HCQ up to 5 μs. A negative absorption peak at 320 nm arises due to the strong parent HCQ absorption at 330 and 342 nm. The experimentally obtained transient absorption spectra were corrected by taking care of the parent absorption using equation (2) [Rath and Mukherjee, 1997].

$${\epsilon }_{\mathrm{R}}={\epsilon }_{\mathrm{P}}+\frac{\Delta {\mathrm{O}\mathrm{D}}_{\mathrm{O}\mathrm{b}\mathrm{s}}{\mathrm{G}}_{\mathrm{d}}{\epsilon }_{\mathrm{d}}}{\Delta {\mathrm{O}\mathrm{D}}_{\mathrm{d}}{\mathrm{G}}_{\mathrm{R}}}$$ (2)

In this equation, ε_R, ε_P and ε_d are the molar extinction coefficient of transient radical, parent molecules and the (SCN)₂ ^•- respectively. ΔOD_obs and ΔOD_d are the observed absorbance values for the transient radical and the (SCN)₂ ^•-, respectively. G_R and G_d are the radical yields for the transient radical and (SCN)₂ ^•- respectively. The ε_d value at 475 nm 8652 mol⁻¹dm³cm⁻¹ and G_d value 2.9 (molecules/100 eV energy absorbed) were used for the calculation of ε_R [Adams et al., 1965]. The corrected transient absorption spectra are presented in Fig. 3b. It shows a peak at 330–340 nm with a shoulder at around 400 nm and the absorption extends up to 550 nm.

Fig. 3.

(a) Transient absorption spectra measured in the case of the reaction of HCQ with ^•OH radicals up to 5 μs time, Inset: Kinetic profiles obtained at 330 nm and 400 nm in the short time scale, (b) Corrected transient absorption spectra obtained after applying the correction formula for the same observed transient spectra.

The kinetic profiles obtained at 330 and 400 nm at two different time scales are shown in inset of Fig. 3a for short scale. Long-time scale decay kinetic profiles are shown in Fig. S1 in the supplementary information. It indicates that the reaction between HCQ and ^•OH radicals lead to the formation of transient intermediate species having two absorptions at 330–340 nm (where the parent molecule has a strong absorption, parent bleaching) and at around 400 nm. The formation profiles at 400 nm are best fitted with the pseudo-first order kinetics and the results are listed in Table 1 . From these studies it is clear that two different types of transient intermediate species are formed by the reaction with ^•OH radicals. One having a strong absorption (along with parent bleaching) in the range 330–340 nm, could be assigned due to the formation of [HCQ⁺] cation and the other having absorption at around 400 nm could be due to the formation of an adduct [HCQ: ^•OH]. Both the cation as well as the adduct have a similar formation rate constant, however, the adduct decays faster as compared to the cation. Nevertheless, both transient intermediate species lead to the degradation of parent HCQ molecule resulting in to a decrease of its absorbance value at the peak position 342 nm, which will be discussed in the next section. A most probable rection mechanism has been given below.

$$\mathrm{H}\mathrm{C}\mathrm{Q}{+}^{{}_{•}}{\mathrm{O}}^{}\mathrm{H}\stackrel{{\mathrm{k}}_{\mathrm{f}}=9.5\mathrm{x}{10}^9{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}}{\to }\begin{array}{c}\left[\mathrm{H}\mathrm{C}{\mathrm{Q}}^{+}\right]\\ 330\mathrm{n}\mathrm{m}\end{array}\stackrel{{\mathrm{k}}_{\mathrm{d}}=\mathrm{s}\mathrm{l}\mathrm{o}\mathrm{w}}{\to }\mathrm{O}\mathrm{x}\mathrm{i}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (3)

$$\mathrm{H}\mathrm{C}\mathrm{Q}{{+}_{}}^{•}\mathrm{OH}\stackrel{{\mathrm{k}}_{\mathrm{f}}=9.5\mathrm{x}{10}^9{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}}{\to }\begin{array}{c}\left[\mathrm{H}\mathrm{C}\mathrm{Q}{:}^{{}_{•}}\mathrm{OH}\right]\\ 400\mathrm{n}\mathrm{m}\end{array}\stackrel{{\mathrm{k}}_{\mathrm{d}}=1.1\mathrm{x}{10}^4{\mathrm{s}}^{-1}}{\to }\mathrm{O}\mathrm{x}\mathrm{i}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (4)

Table 1.

Kinetic parameters obtained from the pulse radiolysis studies for the reactions of e_aq⁻ and ^•OH radicals with HCQ at 400 nm.

HCQ + eaq-		HCQ + •OH
kf (M−1 s−1)	kd (s−1)	kf (M−1 s−1)	kd (s−1)
2.0 x 109	4.2 x 103	9.5 x 109	1.1 x 104

Similarly, the transient absorption spectra recorded for the reaction between HCQ and e_aq ⁻ at short time scale up to 10 μs are shown in Fig. 4 a. The negative absorption seen in this figure is due to the strong parent absorption (bleaching signal) in that region. The observed transient absorption spectra were corrected using the already explained procedure in the case of ^•OH radical reaction, and is shown in Fig. 4b. It is clear that the transient intermediate species formed in this case has a very strong absorption at 330 nm, a weak absorption at around 400 nm and a very broad absorption with a relatively low intensity peak at around 640 nm.

Fig. 4.

(a) Transient absorption spectra measured in the case of the reaction of HCQ with e_aq⁻ up to 10 μs time, Inset: Kinetic profiles obtained at 330 nm, 400 and 640 nm in the short time scale, (b) Corrected transient absorption spectra obtained after applying the correction formula for the same observed transient spectra.

The kinetic profiles obtained at 330, 400 and 640 nm at short time scales are shown in inset of Fig. 4a. Long-time scale decay kinetic profiles are shown in Fig. S2 in the supplementary information. The formation profiles are best fitted with the pseudo-first order kinetics and the results are listed in Table 1. As is seen here, three different transient intermediate species are formed by the reaction with e_aq ⁻. One having a strong absorption around 330 nm (with strong parent bleaching signal) could be due to the formation of a relatively short-lived [HCQ⁻] anion, another having a weak absorption could be due to the formation of an adduct [HCQ:e_aq ⁻] and the other having a very broad absorption with a peak at around 640 nm could be due to a relatively long-lived species other than the above two. At 640 nm, absorption due to hydrated electron is apparent, from the fast formation signal followed by its fast decay within a few micro seconds. However, this signal does not come back to zero, indicating there could be a simultaneous formation of some new transient intermediate species, which has a broad absorption band with peak at around 640 nm (Fig. 4a). The bleach signal at 330 nm was found to decay with a first order rate constant 1.0 x 10⁴ s⁻¹, which is an order of magnitude higher as compared to those of other two transient species having absorption at 400 and 640 nm. Nevertheless, all these transient intermediate species lead to the degradation of HCQ leading to a decrease in its absorbance value at the peak position 342 nm, which will be discussed in the next section. A most probable rection mechanism has been given below.

$$\mathrm{H}\mathrm{C}\mathrm{Q}+{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-}\stackrel{{\mathrm{k}}_{\mathrm{f}}=2.0\mathrm{x}{10}^9{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}}{\to }\begin{array}{c}\left[\mathrm{H}\mathrm{C}{\mathrm{Q}}^{-}\right]\\ 330\mathrm{n}\mathrm{m}\end{array}\stackrel{{\mathrm{k}}_{\mathrm{d}}=1.0\mathrm{x}{10}^4{\mathrm{s}}^{-1}}{\to }\mathrm{R}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (5)

$$\mathrm{H}\mathrm{C}\mathrm{Q}+{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-}\stackrel{{\mathrm{k}}_{\mathrm{f}}=2.0\mathrm{x}{10}^9{\mathrm{M}}^{-1}{\mathrm{s}}^{-1}}{\to }\begin{array}{c}\left[\mathrm{H}\mathrm{C}\mathrm{Q}:{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-}\right]\\ 400\mathrm{n}\mathrm{m}\end{array}\stackrel{{\mathrm{k}}_{\mathrm{d}}=4.2\mathrm{x}{10}^3{\mathrm{s}}^{-1}}{\to }\mathrm{R}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (6)

$$\mathrm{H}\mathrm{C}\mathrm{Q}+{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-}\to \begin{array}{c}\left[\mathrm{H}\mathrm{C}{\mathrm{Q}}^{.-}\right]\\ 640\mathrm{n}\mathrm{m}\end{array}\stackrel{{\mathrm{k}}_{\mathrm{d}}=6.6\mathrm{x}{10}^3{\mathrm{s}}^{-1}}{\to }\mathrm{R}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (7)

3.2. Degradation studies

The degradation of HCQ in aqueous media due to the reaction with free radicals is very important due to its availability towards various biological activities. Therefore, its degradation studies have been carried out with important radicals, namely, e_aq ⁻, ^•OH and O₂ ^•- in aqueous solution at the natural pH. The yield of H₂O₂ is very low in water radiolysis under electron beam irradiation, and therefore was not considered for its contribution towards the oxidative degradation of HCQ.

In order to maintain an oxidizing environment with ^•OH radicals, the solution is saturated with N₂O gas, which reacts with e_aq ⁻ and produces ^•OH radicals with a very high-rate constant k = 9.1 x 10⁹ M⁻¹ s⁻¹) [Janata et al., 2002].

$${\mathrm{N}}_2\mathrm{O}+{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-}\stackrel{\mathrm{H}2\mathrm{O}}{\to }{\mathrm{N}}_2+{\mathrm{O}\mathrm{H}}^{-}+{}^{•}\mathrm{O}\mathrm{H}$$ (8)

Similarly, for the investigation of reaction through e_aq ⁻ in a reducing environment in the solution, 0.1 M tert-butanol (t-BuOH) is added in order to quench ^•OH radicals (also H atom) under a de-aerated condition by purging with inert gas like nitrogen or Argon. The t-BuOH^• radical (^•CH₂C(CH₃)₂OH) formed in the reaction is very less reactive and not expected to interfere in the reaction [Buxton et al., 1988].

$${{\left({\mathrm{C}\mathrm{H}}_3\right)}_{3}COH+{}^{•}\mathrm{O}H/{\mathrm{H}}^{•}\to {}^{•}\mathrm{C}{\mathrm{H}}_{2}C\left({\mathrm{C}\mathrm{H}}_3\right)}_2\mathrm{O}\mathrm{H}+{\mathrm{H}}_2\mathrm{O}/{\mathrm{H}}_2$$ (9)

N₂O saturated aqueous solution was used for the investigation of the reaction of ^•OH radicals with HCQ. The absorption spectra were recorded after each irradiation in the solution kept inside a tightly sealed quartz cuvette. It was observed that there was a systematic reduction in the absorbance at both the peaks (Fig. 5 ) and a minor increase in the absorbance values beyond these two peak positions with increase in the absorbed dose. The absorbance values at the peak position 342 nm were plotted against the absorbed dose, and were found to decrease in a non-linear manner with respect to absorbed dose (Fig. 7). This study clearly indicates that HCQ undergoes degradation through the reaction with ^•OH radicals. The pH of the final degraded solution was found to be 4.2 which indicates that the solution becomes slightly acidic upon oxidative degradation.

$$\mathrm{H}\mathrm{C}\mathrm{Q}+{}^{•}\mathrm{O}\mathrm{H}\to \mathrm{O}\mathrm{x}\mathrm{i}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (10)

Fig. 5.

Absorption spectra recorded after irradiation with different absorbed doses in the N₂O saturated aqueous solution of 1 x 10⁻⁴ M HCQ. Insert: Plot of absorbance at 342 nm vs absorbed dose in kGy.

Fig. 7.

Plots of absorbance at 342 nm vs absorbed dose in kGy for the degradation studies with ^•OH or e_aq⁻.

The oxidative degradation products do not have absorption near the peak position 342 nm but have a weak absorption beyond this peak and thereby giving rise to the appearance of a pale yellow colour in the degraded sample as seen from the absorption spectra. The value of G_(-HCQ) was calculated from the absorption spectra and was found to be 0.036 μmol dm⁻³/J.

N₂ purged aqueous solution containing HCQ and tert-butanol was used for the investigation of the reaction of hydrated electrons, e_aq ⁻, with HCQ. It was observed that there was a systematic reduction in the absorbance at both the peaks (Fig. 6 ) and a minor increase in the absorbance values beyond these two peak positions with increase in the absorbed dose, similar to the case of ^•OH radical reaction. The absorbance values at the peak position 342 nm were plotted against the absorbed dose, and were found to decrease in a similar non-linear manner with respect to absorbed dose (Fig. 7). This study clearly indicates that HCQ also undergo degradation through the reaction with e_aq ⁻. The pH of the final degraded solution was found to be 5.1 which indicates that the solution becomes slightly acidic upon oxidative degradation.

$$\mathrm{H}\mathrm{C}\mathrm{Q}+{\mathrm{e}}_{\mathrm{a}\mathrm{q}}^{-}\to \mathrm{R}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{t}\mathrm{i}\mathrm{v}\mathrm{e}\mathrm{d}\mathrm{e}\mathrm{g}\mathrm{r}\mathrm{a}\mathrm{d}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}$$ (11)

Fig. 6.

Absorption spectra recorded after irradiation with different absorbed doses in the N₂ purged aqueous solution containing 1 x 10⁻⁴ M HCQ and 1% (v/v) tert-butanol. Insert: Plot of absorbance at 342 nm vs absorbed dose in kGy.

The reductive degradation products also do not have absorption near the peak position 342 nm but have a very weak absorption beyond this peak as seen from the absorption spectra. The value of G_(-HCQ) was calculated from the absorption spectra and was found to be 0.024 μmol dm⁻³/J. Similarly, the degradation of HCQ was also investigated with superoxide radicals, O₂ ^•-, in solution containing HCQ, tert-butanol saturated with Oxygen. In this case, e_aq ⁻ reacts with O₂ to produce O₂ ^•- radical with rate constant, 1.9 x 10¹⁰ M⁻¹s⁻¹. Eventually, these O₂ ^•- radicals react with HCQ to undergo subsequent degradation. However, it was observed that the degradation was slower was compared to those observed in the case of ^•OH and e_aq ⁻ cases (Fig. 7).

Survival percentage of HCQ was calculated from the above figure, and was plotted against the concentration of free radicals (Fig. S3 in supplementary information). The survival percentage here was defined as the percentage of HCQ left behind in the solution after the reaction with ^•OH radicals, which was calculated from the absorbance values at 342 nm. The concentration of free radicals was estimated from the absorbed dose and the radiation chemical yield of respective free radicals. It was observed from this figure that the degradation of HCQ is almost identical in both the cases of ^•OH and e_aq ⁻, however, it was slow in the case of O₂ ^•-. This indicates that HCQ undergoes both oxidative as well as reductive degradation in aqueous solution. Nearly 70% degradation takes place with a free radical (^•OH and e_aq ⁻) concentration of about 1 x 10⁻³ M. As the oxidative stress releases ^•OH radicals under physiological conditions [Srivastava and Kumar, 2015; Lipinski, 2011], the further studies related to the degradation was carried out with ^•OH radicals only. It is to be mentioned here that degradation products, similar to those reported by Saini and Bansal, would also be formed in the present study processes [Saini and Bansal, 2013]. However, the degradation products formed in the present study have not been characterized, because the main idea of this study is to understand the reaction pathway and reactivity pattern of free radicals with HCQ.

The oxidative degradation studies have been carried out in the presence of different biologically available compounds, such as ascorbic acid (AA), oxalic acid (OA), citric acid (CA) or gallic acid (GA). These are commonly available compounds in the foods that we consume everyday such as fruits, vegetables and beverages. Therefore, chosen for the investigation of their role in protecting HCQ under oxidative condition. It was found that ascorbic acid has no contribution in the absorbance at 342 nm as seen from the absorption spectra (Fig. S4a – S4c in the supplementary information). The plots of absorbance at 342 nm vs absorbed dose (Fig. 8 a) in the presence of ascorbic acid (from 1 x 10⁻³ M to 5 x 10⁻³ M) indicate that there is a decrease in the oxidative degradation of HCQ under this condition. Therefore, it is expected that the presence of ascorbic acid will certainly protect HCQ from the oxidative degradation. The improved survival percentage of HCQ under an oxidative condition (see Fig. 9 ) supports the above statement. This effect could be predicted from the reaction rate constants of ^•OH radicals with ascorbic acid (1.0 x 10¹⁰ M⁻¹s⁻¹, Buxton et al., 1988) and HCQ (9.5 x 10⁹ M⁻¹s⁻¹, this study). From these rate constants, it is clear that ascorbic acid will be a good candidate for protecting HCQ against oxidative degradation, and which has been experimentally observed in this study.

Fig. 8.

Plots of absorbance at 342 nm vs absorbed dose in kGy for the degradation studies with ^•OH radical in the presence of different concentrations of (a) ascorbic acid (1, 2, and 5 mM), (b) oxalic acid (1, and 5 mM), (c) citric acid (1, and 5 mM)., (d) gallic acid (1, and 2 mM).

Fig. 9.

Plots of % survival of HCQ vs absorbed dose in kGy for the degradation studies with ^•OH radical in the presence of ascorbic acid or gallic acid of 1 mM each.

However, in the presence of oxalic acid, substantial change in the oxidative degradation of HCQ was not noticed up to a concentration of 5 x 10⁻³ M oxalic acid (Fig. 8b) and the degradation results are monitored only up to a lower absorbed dose of 0.6 kGy instead of 2 kGy. However, oxalic acid also does not contribute in the absorbance value at 342 nm (Fig S5a – S5b in the supplementary information). A similar observation was also obtained in the case of citric acid (Fig. 8c). Citric acid too does not have any absorbance at 342 nm (Fig S6a – S6b in the supplementary information). It is to be mentioned here that the rate constants of ^•OH radicals with oxalic acid (1.4 x 10⁶ M⁻¹ s⁻¹, Buxton et al., 1988) and citric acid (5.0 x 10⁷ M⁻¹ s⁻¹, Buxton et al., 1988) are very much lower as compared to that of HCQ. Hence, based on the competition kinetics, it is expected that these two compounds do not have significant ability to protect HCQ from its oxidative degradation.

However, in the presence of gallic acid, a systematic decrease in the oxidative degradation of HCQ with an increase in the concentration of gallic acid up to 2 x 10⁻³ M (Fig. 8d) was found. As in case of other organic acids, gallic acid also does not contribute to the absorbance at 342 nm (Fig S7a and Fig S7b in the supplementary information). The effectiveness of gallic acid towards the protection of oxidative degradation of HCQ could also be supported from its higher reaction rate constant (2.8 x 10¹⁰ M⁻¹ s⁻¹, Marino et al., 2014) with ^•OH radicals as compared to that in the case of HCQ. It was observed from these experiments that the effectiveness of these biomolecules towards the protection of HCQ from oxidative degradation follows the order: gallic acid > ascorbic acid > citric acid > oxalic acid. This was substantiated from the plots of survival percentage values against the absorbed dose (Fig. 9) as well as the plots of survival percentage values against free radical concentrations (Fig S8 in the supplementary information). It is mainly due to their structural parameters, for example, gallic acid has an aromatic benzene ring with three hydroxyl and one carboxyl acid group, ascorbic acid has a five membered ring structure with four hydroxyl groups in all, whereas, citric acid and oxalic acid are aliphatic in nature.

The computational study can also be exploited to characterize the site-specific reactivity parameter towards the electrophilic and nucleophilic attack namely for ^•OH radical and hydrated electron (e_aq ⁻). As discussed in methodology section, condensed Fukui functions were evaluated for this purpose for neutral, mono and di-protonated form of the HCQ for each atom. The graph for such study is shown in Fig. 10 . From graph it is evident that the side chain shows no reactivity towards either electrophilic or nucleophilic attack. The quinoline ring only shows such reactivity trend. From this figure it is evident that all forms of HCQ, whether it is neutral, mono or di-protonated, show similar reactivity trend towards the electrophilic and nucleophilic attack. The carbon atom at 3rd position (C3) shows a maximum reactivity towards an electrophilic attack. In addition, two carbon atoms at 6th and 8th position also shows reactivity towards the ^•OH radical reaction. Nitrogen atoms at 11th position attached to the 4th carbon atom also shows such trend. However, for the electrophilic/nucleophilic attack, we are only considering the carbon atoms. Similarly, for nucleophilic attack, the carbon atom at the 5th position shows a maximum reactivity apart from 2nd and 8th position, which also show similar reactivity. The present theoretical study also evaluates various global reactivity parameters for all the three species of HCQ. These parameters can be seen in Table 2 . It should be noted that a higher basis set, namely, 6–311+g(d) was used for evaluating these parameters. The ^•OH radical reactions with quinoline have been reported in the literature, where it has shown that the ^•OH adducts are the major intermediates formed via both the aromatic rings [Nicolaescu et al., 2003; Nicolaescu et al., 2005]. Based on our studies, possible reaction pathways of the e_aq ⁻ and ^•OH radicals with HCQ molecule could be represented as shown in Scheme 2 .

Fig. 10.

(a) Condensed Fukui index (f^–)for various atom in different species of HCQ for electrophilic attack. (b) Condensed Fukui index (f + )for various atom in different species of HCQ for nucleophilic attack.

Table 2.

Theoretically determined various global reactivity parameters for all the three species of HCQ.

	Global Reactivity Parameters	HCQ (Neutral)	HCQH+ (Monoprotonated)	HCQH2++ (Diprotonated)
1	Ionization potential (I)	5.51	5.66	6.78
2	Electron affinity (EA)	1.40	1.50	2.53
3	Chemical hardness (η = (I–EA)/2)	2.06	2.08	2.13
4	Chemical softness (S = η/2)	1.03	1.04	1.06
5	Chemical potential (μ = –(I + EA)/2)	−3.46	−3.58	−4.66
6	Electronegativity (-μ)	3.46	3.58	4.66
7	Electrophilicity index (ω = μ2/2η)	2.91	3.08	5.10
8	Maximum charge transfer index (ΔNmax) =(-μ/η)	1.68	1.72	2.19

Scheme 2.

Possible reaction pathways showing transient intermediate species with the reactions of e_aq⁻ and ^•OH radicals with HCQ.

4. Conclusions

Hydroxychloquine sulfate readily reacts with e_aq ⁻ and ^•OH free radicals. The formation of transient species with absorption peak at about 400 nm in the case of e_aq ⁻, ^•OH, takes place with rate constants, 2.0 x 10⁹ M⁻¹s⁻¹ and 9.5 x 10⁹ M⁻¹s⁻¹ respectively. However, these transient species decay with slower rate constant as compared to their formation. Hydroxychloquine sulfate undergoes both oxidative as well reductive degradation by reacting with the free radicals ^•OH and e_aq ⁻ respectively. Nevertheless, its degradation could be minimized in the presence of biologically available compounds like ascorbic acid or gallic acid. The effect of various biomolecules towards protecting its degradation was found the in the order, gallic acid > ascorbic acid > citric acid > oxalic acid. It is now clearly evident that the presence of such biomolecules could enhance the stability of this drug molecule in an environment where there would be an unintentional generation of oxidative free radicals like ^•OH. This study is expected to widen the understanding the drugs using hydroxychoroquine sulfate as an ingredient. In addition, computational studies were carried out to determine various global and local reactivity parameters for different forms of HCQ. Site specific reactivity for nucleophilic and electrophilic addition reactions were studied using various computational methods.

Credit author statement

M. C. Rath: Idea, data Formal analysis and interpretation, manuscript writing. S. J. Keny: Experiment, data Formal analysis and manuscript writing. Hari P. Upadhyaya: Computational work, data Formal analysis, interpretation and manuscript writing. S. Adhikari: Data Formal analysis and interpretation, manuscript writing

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Handling Editor: Dr. Jay Laverne

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.