Kinetic Analysis of the Reduction Processes of a Cisplatin Pt(IV) Prodrug by Mesna, Thioglycolic Acid, and Thiolactic Acid

Although Mesna is an FDA-approved chemotherapeutic adjuvant and an antioxidant based largely on its antioxidative properties, kinetic and mechanistic studies of its redox reactions are limited. A kinetic analysis of the reduction processes of cis-diamminetetrachloroplatinum(IV) (cis-[Pt(NH3)2Cl4], a cisplatin Pt(IV) prodrug) by thiol-containing compounds Mesna, thioglycolic acid (TGA), and DL-thiolactic acid (TLA) was carried out in this work at 25.0°C and 1.0 M ionic strength. The reduction processes were followed under pseudo-first-order conditions and were found to strictly obey overall second-order kinetics; the observed second-order rate constant k′ versus pH profiles were established in a wide pH range. A general reaction stoichiometry of Δ[Pt(IV)] : Δ[Thiol]tot = 1 : 2 was revealed for all the thiols; the thiols were oxidized to their corresponding disulfides which were identified by mass spectrometry. Reaction mechanisms are proposed which involves all the prololytic species of the thiols attacking the Pt(IV) prodrug in parallel, designating as the rate-determining steps. Transient species chlorothiol and/or chlorothiolate are formed in these steps; for each particular thiol, these transient species can be trapped rapidly by another thiol molecule which is in excess in the reaction mixture, giving rise to a disulfide as the oxidation product. The rate constants of the rate-determining steps were elucidated, revealing reactivity enhancements of (1.4–8.9) × 105 times when the thiols become thiolates. The species versus pH and reactivity of species versus pH distribution diagrams were constructed, demonstrating that the species ‒SCH2CH2SO3‒ of Mesna largely governs the total reactivity when pH > 5; in contrast, the form of Mesna per se (mainly as HSCH2CH2SO3‒) makes a negligible contribution. In addition, a well-determined dissociation constant for the Mesna thiol group (pKa2 = 8.85 ± 0.05 at 25.0°C and μ = 1.0 M) is offered in this work, which was determined by both kinetic approach and spectrophotometic titration method.


Introduction
Mesna (namely, sodium 2-mercaptoethanesulfonate) is an FDA-approved drug which has been used to reduce the risk of hemorrhagic cystitis (a condition that causes inflammation of the bladder and can result in serious bleeding) in people who receive ifosfamide or cyclophosphamide for cancer treatments [1][2][3]. ese two anticancer drugs in vivo may be converted to urotoxic metabolites such as acrolein.
e protecting mechanism of Mesna is assisting to detoxify these metabolites by reaction of its thiol group with the α,β-unsaturated carbonyl containing compounds including acrolein. us, Mesna is a chemotherapeutic adjuvant [1][2][3]. Moreover, it is also an antioxidant, and its FDA approval was based largely on its antioxidative properties [1][2][3]. On the other hand, the protonated form of Mesna is 2-mercaptoethanesulfonic acid which bears another name: coenzyme M [4][5][6][7]. Coenzyme M is the smallest known organic cofactor and serves as a methyl group carrier in key reactions within the pathway of methane formation from C1 precursors [6]. In the alkene metabolism pathway, it is involved in aliphatic epoxide carboxylation [4][5][6][7]. Cisplatin (cis-[Pt(NH 3 ) 2 Cl 2 ]), the first Pt(II)-based antineoplastic drug, has played a central role in cancer chemotherapies covering a relatively wide spectrum including testicular, ovarian, cervical, breast, bladder, head and neck, esophageal and lung cancers, mesothelioma, brain tumors, and neuroblastoma [8][9][10]. Despite of its tremendous success, it has several severe side effects such as nephrotoxicity, ototoxicity, neurotoxicity, and gastrointestinal toxicity in addition to the acquired resistance to the drug [9,10]. Two general approaches have been developed in order to overcome or minimize these detrimental side effects [11]: (a) the search of adjuvant or rescuing agents which are used together with cisplatin [12][13][14], and (b) the conversation of cisplatin to its Pt(IV) prodrugs by assumption that these prodrugs can be delivered to the tumor sites more efficiently [15][16][17][18]. e first approach has had a very limited success, as such, Mesna has been found to ameliorate significantly the severe side effects of cisplatin in certain types of cancers [19][20][21][22][23]. Furthermore, the combined use of cisplatin, ifosfamide, and Mesna has been exploited to the treatment of advanced ovarian carcinoma [24][25][26]. Whereas the second approach still remains elusive although large efforts have been attempted [11]. Also, it was shown that cisplatin Pt(IV) prodrugs can induce oxidative stress [18]; this might be not surprising since Pt(IV) prodrugs per se are oxidants.
In contrast to the rich biomedical studies of Mesna, the kinetic and mechanistic aspect related to the redox reactions is scanty [27][28][29], which encompasses the interaction of Mesna with cisplatin [27] and the reduction reactions of bromate and bromine [28] and of the anticancer prodrug ruthenium(III) compound [29] by Mesna. We carried out a kinetic analysis of the reduction processes of cis-diamminetetrachloroplatinum(IV) (cis-[Pt(NH 3 ) 2 Cl 4 ]) by Mesna and its structurally related thiols thioglycolic acid (TGA) and DL-thiolactic acid (TLA), cf. their structures in Figure 1. cis-[Pt(NH 3 ) 2 Cl 4 ] per se is anticancer active [30,31] and is also a prodrug of cisplatin since it is usually reduced to cisplatin. e essentially substitution-inertness property of cis-[Pt(NH 3 ) 2 Cl 4 ] enabled us to study its redox reactions in a wide pH range and to derive the reactivity of each protolytic species of reductants [32,33]. We thus divulge the results from the kinetic analysis.

2.2.
Instruments. UV-Vis spectra, absorption measurements, and time-resolved spectra were recorded on a TU-1900 spectrophotometer (Beijing Persee, Inc., Beijing, China) using 1.00 cm quartz cells. Kinetic measurements and rapid scan spectra were carried out on an Applied Photophysics SX-20 stopped-flow spectrometer (Applied Photophysics Ltd., Leatherhead, U.K.). Both spectrophotometer and spectrometer were equipped with a water bath circulation from a thermostat (Lauda Alpha RA8, Belran, NJ, USA); temperature could be controlled to ± 0.1°C. Mass spectra were recorded on an Agilent 1200/6310 ion trap mass spectrometer with electrospray ionization (ESI). An Accumet Basic AB15 Plus pH meter, equipped with an Accumet AccutupH ® combination pH electrode (Fisher Scientific, Pittsburgh, PA), was used to measure the pH values of buffer solutions. Standard buffers of pH 4.00, 7.00, and 10.00, also from Fisher Scientific, were used for calibrations of the electrode just before pH measurements.

Buffer Solutions.
Buffering pairs of H 3 PO 4 /NaH 2 PO 4, HAc/NaAc, NaH 2 PO 4 /Na 2 HPO 4, NaHCO 3 /Na 2 CO 3 , and Na 2 HPO 4 /Na 3 PO 4 (0.15-0.2 M) were prepared to cover a pH range from 2.47 ≤ pH ≤ 11.97. e buffers contained 0.10 M of NaCl and 2.0 mM of EDTA; sodium perchlorate was used to adjust the ion strength (μ) of buffers to an ionic strength of 1.0 M. Details for the buffer preparations are delineated in our earlier works [32,33]. e buffer pairs are listed in Table S1 in the Supplementary Materials (SM).

UV-Vis and Rapid-Scan Spectra.
e UV-Vis spectra for the compounds employed in this work were recorded by use of their solutions freshly prepared in a buffer of pH 4.  Figure S1. In addition, the UV-Vis spectra for the fresh solutions of 0.20 mM cis-[Pt(NH 3 ) 2 Cl 4 ] in buffers of pH 4.42, 6.67, and 10.98 were also recorded, and are given in the lower panel of Figure S1.
In order to gain some insights into these reduction processes, the reduction of cis-[Pt(NH 3 ) 2 Cl 4 ] by Mesna in a buffer of pH 6.25 was probed by recording the rapid scan spectra under a set of reaction conditions. e spectra were recorded between 200-350 nm and are shown in Figure 2(a). Clearly, the absorption band around 236 nm and absorption shoulder round 265 nm decreased concertedly as the reaction proceeded, and no new absorption bands emerged. e absorbance readings at 236 and 266 nm as a function of time are shown in Figure 2(b) (data points). e absorbance versus time curves or kinetic traces were fit by equation (1) [34], where k obsd represents of pseudo-first-order rate constant and where A t , A 0 , and A ∞ pertain to absorbance at time t, zero, and infinity, respectively. e resulting fittings are excellent ( Figure 2(b)) and moreover, the value of k obsd obtained at 236 nm equals to that acquired at 266 nm within the experimental errors. ese kinetic attributes indicate that the reduction process is first-order in [Pt(IV)] and that the absorbance decrease in Figure 2 corresponds to the reduction of cis-[Pt(NH 3 ) 2 Cl 4 ] without other complications. As a matter of fact, the kinetic attributes are in good agreement with the nature of substitution inertness of Pt(IV) complexes in general [35,36]. In the time-resolved spectra recorded for the reduction process of cis-[Pt(NH 3 ) 2 Cl 4 ] by TLA in an acidic medium, similar spectral and kinetic attributes were observed, cf., Figure S2 in the SM.

Second-Order Kinetics and Data
Collection. e effects of varying [thiol] tot on the reduction rates were studied, aiming at ascertaining the reaction order of the thiols. In most of buffer solutions, [thiol] tot was varied from 0.20 to 2.00 mM and the reaction medium had big enough buffering capacities so that the variation of [thiol] tot did not cause any pH changes in the medium. e kinetic traces were followed by the stopped-flow spectrometer in a wavelength region from 265 to 280 nm. Pseudo-first-order rate constant k obsd was acquired from the simulations of kinetic traces by equation (1). For each [thiol] tot , 3-5 duplicate runs were made, and the values of k obsd were the averages from the duplicated runs. Standard deviations are usually much less than 5%.
Plots of k obsd versus [thiol] tot in the case of Mesna are shown in Figure 3  us, the reduction reactions are first-order in [thiol] tot ; thus, an overall second-order rate law (2) is warranted, where k′ denotes the observed second-order rate constant and k′ � k obsd /[thiol] tot : We collected a large body of kinetic data at 25.0°C and μ � 1.0 M covering a wide pH range. Values of k′ were calculated from the linear k obsd versus [thiol] tot plots for all the three thiols and are summarized in Tables S1-S3 in the SM. Alternatively, logk′ versus pH plots are illustrated in Figures 4 and 5 (data points).

Reaction Stoichiometry and the Oxidation Products.
A spectrophotometric titration method was employed to find the stoichiometry for the redox reactions between cis-[Pt(NH 3 ) 2 Cl 4 ] and the thiols [32][33][34].
e plots of absorption values at 265 nm versus the ratio [thiol] tot /[Pt(IV)] are displayed in Figure 6. For each thiol, the data points clearly follow two crossing straight lines affording an intersection point. e stoichiometric ratios were acquired ESI mass spectra were recorded for reaction mixtures of 1 mM Pt(IV) with 8 mM TGA and of 1.0 mM Pt(IV) with 8 mM TLA in 10 mM HCl after a reaction time about 1 h; the obtained spectra are shown in Figures S5 and S6 in the SM together with the peak assignments. From the peak assignments, formations of disulfides HOOCCH 2 S-SCH 2 COOH and HOOCCMeS-SCMeCOOH were confirmed for the reactions of TGA and TLA, respectively.

Reaction Mechanisms.
e observed second-order rate constants k′ increase several orders of magnitude (Table S1- [35,36]. In contrast, the reduction processes of cis-[Pt(NH 3 ) 2 Cl 4 ] by the three thiols are much quicker, thus ruling out the possibility that the reduction reactions proceed via ligand substitution(s). is is consistent with the attributes found in the time-resolved spectra discussed above. If all the protolytic species of TGA, TLA, and Mesna are logically assumed to be able to reduce the Pt(IV) complex, the reaction mechanism delineated in Scheme 1 is suggested for the reactions of TGA and TLA [32,33]. e reaction mechanism described in Scheme 2 is proposed analogously for the Mesna reaction. In the mechanisms, the reactions designated by k 1 -k 3 are the rate-determining steps. Each of the steps is taking place via an attack on one of the two axially-coordinated chlorides by the sulfur atom of the thiols, forming chloride-bridged transitions states which in the case of Mesna are depicted conceivably as follows [37][38][39]:  In the transition states, partial bond formation (or bridge formation) between Cl and S atom occurs whereas concurrently the bonds of Cl-Pt-Cl are partially broken [37][38][39].
e collapses of the transition states generate transient chlorothiol and/or chlorothiolate species [37][38][39]. For each particular thiol, these transient species can be trapped rapidly by another thiol molecule which is in excess in the reaction mixture, giving rise to a disulfide as the oxidation product [32,33,37].

Rate Constants of the Rate-Determining
Steps.
e reaction mechanisms outlined in Schemes 1 and 2 are very similar and a common rate expression can be derived as follows: Equation (5) is equivalent virtually to equation (2), where a H is the proton activity and corresponds to the measured pH values by a relation: pH � −log(a H ). A comparison of equations (2) and (5) renders

Journal of Chemistry
Equation (6) was utilized to simulate the kinetic data in Figures 4 and 5. In the simulations, the rate constants k 1 -k 3 were unknowns and treated as adjustable parameters. On the other hand, if the acid dissociation constants K a1 and K a2 for the thiol-containing compounds are available at the relevant conditions, they will be used as direct inputs, minimizing the number of adjustable parameters. Fortunately, the acid dissociation constants of TGA were reported to be pK a1 � 3.53 and pK a2 � 10.05 at 25.0°C and μ � 1.0 M [37]. When these pK a values were used direct inputs, equation (6) was employed to simulate the k′-pH dependence data by use of a weighted nonlinear least-squares method. e simulated result for TGA turned out to be good and is shown in Figure 4(a), concurrently providing values for the rate constants of k 1 -k 3 which are listed in Table 1. e acid dissociation constants of TLA were reported as pK a1 � 3.38 and pK a2 � 9.93 at 25.0°C and μ � 0.50 M [40] and were also utilized as direct inputs although the ionic strength is slightly differentiated. e simulation of equation (6) to the k′-pH dependence data revealed an essentially perfect fit, as shown in Figure 4(b). e acquired values of k 1 -k 3 from the simulation are also listed in Table 1.
A reliable value of the thiol dissociation constant for Mesna (i.e. pK a2 in Scheme 2) appears nonavailable in the literature under our experimental conditions except that some articles [2,41] mentioned a value of 9.2 without specifying any conditions, which is in contrast to the fact that Mesna has been subjected to extensive medical studies as illustrated in the introduction section. e pK a1 value of coenzyme M in Scheme 2 was not found in the literature.   (6) to the experimental data by use of a weighted nonlinear least-squares method.  (7) to the experimental data by use of a weighted nonlinear least-squares method.
Due to the biomedical importance of Mesna, acquiring a reliable value of the thiol dissociation constant is appealing. On the other hand, the pK a1 value of coenzyme M is anticipated to be close to that of ethanesulfonic acid (CH 3 CH 2 SO 3 H), which was reported to be pK a � 1.65 at 25.0°C [42].
For the kinetic data analysis by equation (6), an initial value of pK a1 � 1.65 and a tunable pK a2 were tried for the simulations. e trial simulations indicated that the k 1 value was indeterminate when pK a1 was varied from 1.2 to 2.0. us, the k 1term in equation (6) is negligible, leading to follows: Equation (7) was then employed for simulation of the k′-pH dependence data, and the simulated result is shown in Figure  5, conferring well-defined values for pK a2 � 8.86 ± 0.08 and k 3 � (3.29 ± 0.09) × 10 5 M −1 s −1 at 25.0°C and μ � 1.0 M. Furthermore, more simulations were performed with pK a1 values changed from 1.2 to 2.0 (the real pK a1 value is certainly in this region), yielding the robust values for pK a2 and k 3 mentioned above. e obtained k 2 values has a small variation, but only changed from 0.32 ± 0.03 to 0.42 ± 0.03 M −1 s −1 .
e results are given in Table 1 where k 2 value is taken as an average. erefore, although the pK a1 value of coenzyme M is not known, we obtained the well-defined values for pK a2 and k 3 , and a reasonable value for k 2 .

Determination of the iol Dissociation Constant of
Mesna. In order to further ascertain the pK a2 value of Mesna obtained by the above kinetic approach, we thus determined it by the spectrophotometric titration method [43][44][45]. A series of buffer solutions containing 0.12 mM Mesna at 1.0 M ionic strength were prepared covering a pH range from 6.77 to 11.97. ose solutions were flushed by nitrogen gas and concurrently thermoequilibated at 25.0°C for 10 min. For each pH, absorbance at 235 nm was measured with the corresponding buffer without Mesna as a reference. e measured absorption value as a function of pH is given in Figure 7 (data points). Equation (8) is a standard correction between the measured absorbance and pK a2 [43][44][45] in which ε 2 and ε 3 are the molar absorptivities for HSCH 2 CH 2 SO 3 and -SCH 2 CH 2 SO 3 -, Abs(235 nm) � [Mesna] tot ε 3 + ε 2 10 pK a2 − pH ( ) respectively. e ε 2 value was determined separately to be 59.0 ± 0.5 M −1 cm −1 by use of three solutions of ca. 10 mM Mesna between pH 4 and 5. Equation (8) was then employed to simulate the data in Figure 7 using a nonlinear least-squares routine; the simulation resulted in a good fit, furnishing pK a2 � 8.85 ± 0.05 and ε 3 � (5.6 ± 0.1) × 10 3 M −1 cm −1 at 25.0°C and μ � 1.0 M. e excellent agreement between the pK a2 values obtained by the kinetic approach and by the spectreophotometric titration method emphasizes that we provide a reliable value for Mesna thiol dissociation and that the kinetic approach is a good method for acquisition of pK a values [46,47].

Reactivity of Species the ree iols towards the Reduction of cis-[Pt(NH 3 ) 2 Cl 4 ].
Our kinetic data collection in the wide pH range enables us to derive the reactivity of all the protolytic species of the three thiols towards the reduction of cis-[Pt(NH 3 ) 2 Cl 4 ] and to make a comparison. e ratios of k 1 : k 2 : k 3 were found to be 0.038 : 1 : 1.7 × 10 5 for TGA, 0.031 : 1 : 1.4 × 10 5 for TLA, and 0 : 1 : 8.9 × 10 5 for coenzyme M (Mesna), respectively. Clearly, the deprotonation of the carboxylic acids in TGA and TLA gives a reactivity enhancement of about 30 times which is caused by an inductive effect conferred by the deprotonations. When the thiols become thiolates after a further deprotonation, reactivity enhancements of (1.4-8.9) × 10 5 times are engendered for all the thiols. ese huge reactivity enhancements account convincingly for the logk′ versus pH profiles in Figures 4 and 5.
More straightforwardly, the species versus pH and the species reactivity versus pH distribution diagrams were constructed [33,46], which are displayed in Figures S7-S9 in the SM. For all the thiols, species II dominate in the existing forms from pH 7.0 to 8.0 but make negligible contributions to their respect total reactivity. On the other hand, species III play dominant roles in determining the total reactivity whereas their populations are extremely low in the pH region from 7 to 8. In particular, the protolytic species -SCH 2 CH 2 SO 3 largely dominates the total reactivity of Mesna from above pH 5; contrarily, the form of Mesna per se (mainly as HSCH 2 CH 2 SO 3 -) makes a negligible contribution to the total reactivity when pH > 5, cf. Figure S10. It follows that the protolytic species -SCH 2 CH 2 SO 3 but not HSCH 2 CH 2 SO 3 of Mesna may play a leading role in some pharmacological processes of this drug.
Mechanistically, the reduction processes of cis- is can be accounted for in terms of transition state stabilization which is ascribed to the strong σ-donor and π-acceptor properties of the cyanide ligands in [PtCl 2 (CN) 4 ] 2- [48].

Conclusions
e reduction processes of a cisplatin Pt(IV) prodrug cis-[Pt(NH 3 ) 2 Cl 4 ] by Mesna, TGA and TLA strictly follow overall second-order kinetics, and the k′ versus pH profiles have been established in a wide pH range. e proposed reaction mechanisms involve all the prololytic species of the thiols attacking the Pt(IV) in parallel, which are the ratedetermining steps. e rate constants of these rate-determining steps have been elucidated revealing reactivity enhancements of (1.4-8.9) × 10 5 times when the thiols become thiolates. e constructed species versus pH and species reactivity versus pH distribution diagrams demonstrated that the species -SCH 2 CH 2 SO 3 of Mesna largely governs the total reactivity when pH > 5; contrarily, the form of Mesna per se (mainly as HSCH 2 CH 2 SO 3 -) makes a negligible contribution. In addition, a well-determined and reliable dissociation constant for the Mesna thiol group   (pK a2 � 8.85 ± 0.05 at 25.0°C and μ � 1.0 M) is offered in this work.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.

Acknowledgments
Financial support of this work by grants from the Critical R&D Plans of Shandong Province (Major Scientific Innovation Projects, 2019JZZY010516) and from the Natural Science Foundation of Shandong Province (ZR2019MB015) is gratefully acknowledged. e authors thank Dr. Jingran Dong and Mr. Liyao Xu for some experimental assistance.  Figure S1  (HAc/NaAc), pH 6.67 (NaH 2 PO 4 /Na 2 HPO 4 ), and pH 10.98 (Na 2 HPO 4 /Na 3 HPO 4 ). For each spectrum, the corresponding buffer was used as a reference. Figure S2  e solid-curves were the best fits of equation (1) to the experimental data by a nonlinear squares method, affording values of k obsd � (7.7 ± 0.2) × 10 −3 s −1 at 224 nm and k obsd � (8.3 ± 0.2) × 10 −3 s −1 at 270 nm. Figure S3: pseudo-first-order rate constants k obsd versus [TGA] tot in buffer solutions of different pHs at 25.0°C and μ � 1.0 M. Figure S4 Figure S7 (upper panel): TGA species versus pH distribution diagram at 25.0°C which was calculated by use of pK a1 � 3.52 and pK a2 � 10.05. Lower panel: reactivity of the TGA species versus pH distribution diagram in the reduction of cis-[Pt(NH 3 ) 2 Cl 4 ]; the above pK a values and k 1 � 0.26, k 1 � 6.85, and k 4 � 1.15 × 10 6 M −1 s −1 in Table 1 were employed in the calculation. Species I � HSCH 2 COOH; II � HSCH 2 COO -; III � -SCH 2 COO -. Figure S8 (upper panel): TLA species versus pH distribution diagram at 25.0°C which was calculated by use of pK a1 � 3.38 and pK a2 � 9.93. Lower panel: reactivity of the TLA species versus pH distribution diagram in the reduction of cis-[Pt(NH 3 ) 2 Cl 4 ]; the above pK a values and k 1 � 0.19, k 1 � 6.2, and k 4 � 8.8 × 10 8 M −1 s −1 in Table 1 were employed in the calculation. Species I � HSCHMeCOOH; II � HSCHMeCOO -; III � -SCHMeCOO-. Figure S9 (upper panel): species of coenzyme M versus pH distribution diagram at 25.0°C which was calculated by use of pK a1 � 1.65 (assumed) and pK a2 � 8.85. Lower panel: reactivity of the coenzyme M species versus pH distribution diagram in the reduction process of cis-[Pt(NH 3 ) 2 Cl 4 ]; the above pK a values and k 1 � 0, k 1 � 37, and k 4 � 3.29 × 10 5 M −1 s −1 in Table 1